Comprehensive evaluation method based on combination of set pair theory and radar chart method

Abstract

This invention relates to a comprehensive evaluation method based on a combination of set pair theory and radar chart method, comprising: constructing the expected variance type of binary connection numbers for each evaluation index at each observation point, and calculating the modulus and argument of the binary connection numbers; calculating the central angle of each evaluation index in the radar chart according to its weight, and dividing the radar chart into corresponding index regions based on the central angle of each evaluation index; drawing sectors in the corresponding index regions on the radar chart with the modulus of the binary connection numbers as the radius and the argument as the central angle; and evaluating each observation point by the total area and total perimeter of each sector on the radar chart. This invention addresses the shortcomings of existing radar chart methods by using set pair analysis theory to construct binary connection numbers from data statistics, and providing the distribution rules of the modulus and argument of the connection numbers on the radar chart. It uses the sector area and perimeter occupied by the connection numbers as evaluation functions to rank the merits of various evaluation schemes.

Description

Technical Field

[0001] This invention belongs to the field of comprehensive data evaluation technology, specifically involving a comprehensive evaluation method based on a combination of set pair theory and radar chart method. Background Technology

[0002] In modern society, the scale of information is ever-increasing, and its incompleteness and uncertainty are becoming more and more common. Faced with the complex multi-indicator comprehensive evaluation problems caused by massive amounts of ambiguous information, existing information processing tools are inadequate. How to effectively process the large-scale uncertain information obtained using more reasonable tools and derive corresponding evaluation results has become a key point in comprehensive evaluation.

[0003] Many scholars and related institutions have done a lot of work on comprehensive evaluation models (see references [1-8]). Radar chart method is a multivariate comparative analysis technique that can realize the comprehensive evaluation of multiple indicators and is often used in the comprehensive evaluation of various fields. The basic principle of radar chart method is to express the various indicator data of the object to be evaluated in a graphical way, and to realize the comprehensive evaluation of the object to be evaluated by comparing the area and perimeter of the graph. Reference [5] uses Weighting the indicators Converted to the corresponding sector angle Then, the values ​​of each indicator are calibrated at... On the angle bisector, connect each calibration point to form a new radar chart. Quantitative evaluation of the index values ​​of each evaluation point is achieved by comparing the area, perimeter and other characteristic quantities of the radar chart. Reference [6] draws a reverse radar chart for each power quality index. The power quality is evaluated by comparing the total area and total perimeter of the reverse radar chart. Reference [7] replaces the triangular area of ​​the traditional radar chart with a fan-shaped area and selects a new evaluation function to achieve a comprehensive evaluation of the power quality of the power distribution system. Reference [8] allocates the fan-shaped area according to the order of the weights of each index, uses the diagonal of the fan-shaped area as the index axis to draw the radar chart, and uses the area and perimeter of the radar chart as the evaluation function for the final comprehensive evaluation. Although the comprehensive evaluation based on the radar chart method has achieved certain results, there are still shortcomings: (1) With the rapid development of computer and big data technology, the storage and retrieval of massive monitoring data has become a reality. However, the data used in the existing radar chart method is often only data from a certain time segment, and the information value of massive monitoring data has not been fully explored; (2) The division of the sector area of ​​each indicator has failed to effectively solve the problem of the boundary between adjacent indicators; (3) The area and perimeter of the radar chart drawn occupy too much irrelevant area, resulting in a large calculation error and affecting the final evaluation result.

[0004] References:

[0005] [1]MILANOVI JV, ABDELRAHMAN S, LIAO H. Compound index for powerquality evaluation and benchmarking[J]. IET Generation, Transmission &Distribution, 2018, 12(19): 4269-4275.

[0006] [2] YUN JY, CHAO T, YUN TX, et al. A projection pursuit model optimized by free search: for the regional power quality evaluation[J]. TheOpen Cybernetics & Systemics Journal, 2015, 9(1): 1422-1427.

[0007] [3] SHARMA A, RAJPUROHIT BS, SINGH S N. A review on economics ofpower quality: impact, assessment and mitigation[J]. Renewable andSustainable Energy Reviews, 2018, 88: 363-372.

[0008] [4] Zhong H, Li Q. A comprehensive evaluation model of power quality based on blind number and variable fuzzy sets theory[C]. InternationalConference on Integrated Circuits and Microsystems. IEEE, 2016:279-285.

[0009] [5] Li Guodong, Li Gengyin, Yang Xiaodong, et al. Comprehensive power quality assessment model based on radar chart method [J]. Automation of Electric Power Systems, 2010, 34(14): 70-74.

[0010] [6] Liu Ronghui, Wang Yichao. Comprehensive evaluation of power quality based on improved radar chart method and Gaussian membership degree [J]. Electrical Measurement & Instrumentation, 2018, 55(14): 69-74.

[0011] [7] Cheng Zhiyou, Zhu Weiwei, Tao Qing, et al. A method for power quality assessment of power distribution systems based on improved radar charts [J]. Electrical Measurement & Instrumentation, 2019, 56(14): 34-39.

[0012] [8] Qiao Pengcheng, Wu Zhengguo, Li Hui. A comprehensive power quality assessment method based on improved radar chart method [J]. Electric Power Automation Equipment, 2011, 31(6): 88-92. Summary of the Invention

[0013] The purpose of this invention is to provide a comprehensive evaluation method based on the combination of set pair theory and radar chart method. Addressing the shortcomings of existing radar chart methods in processing large-sample measured data, this invention performs statistical analysis on the standardized large-sample data, uses set pair analysis theory to construct binary connection numbers from the data statistics, and provides the distribution rules of the modulus and argument of the connection numbers on the radar chart. The sector area and perimeter occupied by the connection numbers are used as evaluation functions to rank the merits of various evaluation schemes. Finally, the proposed comprehensive evaluation model of the novel radar chart method is applied to an example, demonstrating the feasibility and superiority of this method.

[0014] To achieve the above objectives, the present invention provides the following technical solution:

[0015] A comprehensive evaluation method based on a combination of set pair theory and radar chart method is characterized by:

[0016] Construct the expected variance type binary connection coefficient for each evaluation index at each observation point, and calculate the modulus and argument;

[0017] Calculate the central angle of each evaluation indicator in the radar chart based on its weight, and divide the radar chart into corresponding indicator areas based on the central angle of each evaluation indicator.

[0018] A sector is drawn on the radar chart within the corresponding index area, with the modulus as the radius and the argument as the central angle; the total area and total perimeter of each sector on the radar chart are used to evaluate each observation point.

[0019] Furthermore, the binary link coefficient is an expected variance type binary link coefficient:

[0020] set up and Observation points Indicators measured value The expected value and variance are then

[0021]

[0022] This corresponds to the expected variance type of the binary contact number;

[0023] The expected variance type binary connection coefficient is divided into two cases:

[0024] Scenario 1, when hour, Let the standard expected value variance-coefficient be denoted as . for The model, called for The angle of the argument;

[0025] Scenario 2, when hour, Let Variance be the general expected value-variance correlation coefficient, then for The model, for The angle of the argument;

[0026] according to Given the given value, select the corresponding case and calculate the corresponding modulus and argument.

[0027] Furthermore, the weights of each evaluation indicator are determined. Through formula Calculate the corresponding central angle in the radar chart, and starting from zero degrees, plot the index areas on the radar chart in a counterclockwise direction according to the size of the central angle.

[0028] Furthermore, each corner Convert using the following formula:

[0029]

[0030] Using the starting boundary of each indicator region as Starting from the edge, draw a circle with a radius of modulo 1 in a counterclockwise direction. The central angle is A sector.

[0031] Furthermore, the evaluation of each observation point includes:

[0032] First, compare the total area of ​​each sector at each observation point; the sector with the larger total area is better than the sector with the smaller total area.

[0033] When the total area of ​​the observation points is equal, the total perimeter of each sector at each observation point is compared, and the sector with the larger total perimeter is superior to the sector with the smaller total perimeter.

[0034] When the total perimeter of the observation points is equal, the task is indistinguishable.

[0035] Furthermore, before constructing the expected variance type binary correlation coefficient, the evaluation indicators are standardized: the evaluation indicators are divided into four types: interval type indicators, fixed type indicators, benefit type indicators and cost type indicators, and a standardized model of the evaluation indicators is given.

[0036] Furthermore, the normalized model of the fixed index is as follows:

[0037] For fixed indicators, the relative superiority is:

[0038]

[0039]

[0040] in, For observation point Indicators The measured value, As an indicator The ideal value.

[0041] Furthermore, the normalized model of the interval-type indicator is as follows:

[0042] For interval-type indicators, the relative superiority is:

[0043]

[0044]

[0045] in, For observation point Indicators The measured value, and These are the actual measured values ​​of the indicators. The optimal lower and upper limits, for Maximum deviation These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0046] Furthermore, the standardized model for the cost-type indicator is as follows:

[0047] For cost-related indicators, the relative superiority is:

[0048]

[0049] in, For observation point Indicators The measured value, These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0050] Furthermore, the standardized model for the aforementioned benefit-type indicators is as follows:

[0051] For benefit-oriented indicators, their relative superiority is:

[0052]

[0053] in, For observation point Indicators The measured value, These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0054] To address the shortcomings of existing radar chart methods in the comprehensive evaluation of large sample measured data, the beneficial effects of this invention are as follows:

[0055] (1) The method of the present invention can realize the evaluation of long-term monitoring data. By calculating the expectation and variance of large sample data, the statistical characteristic values ​​of large sample data are obtained. The expected variance correlation coefficient constructed in this way can reflect the true characteristics of each indicator of the observation point on a long time scale, and the final evaluation result will inevitably be more in line with reality;

[0056] (2) The method of the present invention makes a comprehensive judgment by calculating the area and perimeter of the index value sector, which avoids the error caused by the large area and perimeter of irrelevant regions in the index value polygon in the traditional radar chart method.

[0057] (3) In the method of the present invention, the sector area of ​​each statistical value of each indicator at each observation point is determined by the connection number. The value of determines the proportion of data used in the evaluation to the total measured data. Therefore, the appropriate value can be selected based on the accuracy requirements of the evaluation results. value. Attached Figure Description

[0058] Figure 1 This is a schematic diagram of the radar chart method based on set pair theory.

[0059] Figure 2 This is a schematic diagram of the traditional radar chart method. Detailed Implementation

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to specific examples. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0061] Set pair analysis, proposed by Zhao Keqin in 1989, is a systematic mathematical theory for dealing with uncertain problems. The connection number is a key tool in set pair analysis for handling uncertainty. It studies system development from the perspective of the unity of opposites between certainty and uncertainty, and is well-suited for information mining from large sample data. As a primary tool of set pair theory, the connection number, on the one hand, links determinable quantities with their scope; on the other hand, it dynamically connects macroscopically definite quantities with microscopically uncertain quantities.

[0062] This paper designs a comprehensive evaluation method based on the combination of set pair theory and radar chart method (hereinafter referred to as the "connection number radar chart method"). This method combines set pair theory with radar chart method, and after standardizing the monitoring data of each indicator, performs statistical analysis on large sample data. It uses set pair analysis theory to construct binary connection numbers from statistical characteristics, and provides the distribution rules of the modulus and azimuth of the connection numbers on the radar chart. The sector area and perimeter occupied by the connection numbers are used as evaluation functions to rank the merits of each object to be evaluated. The method is mainly divided into the following three parts:

[0063] The first part categorizes evaluation indicators (which can be simply referred to as indicators) into four types: interval indicators, fixed indicators, benefit indicators, and cost indicators, and provides a standardized model for evaluation indicators.

[0064] Most assessments involve a comprehensive evaluation of multiple indicators; therefore, these indicators must be standardized before the assessment can be conducted. Indicator types can be categorized into four types: interval-based indicators, fixed indicators, cost-based indicators, and benefit-based indicators. The standardized models for these four types of indicators are as follows:

[0065] For the subject of evaluation Observation points Select each indicator at each observation point The measured data (measured values) can be selected from long-term data or data from different time periods selected multiple times and then summarized. For indicators with ambiguity, a questionnaire survey scoring method can be used as the measured data.

[0066] 1. For fixed indicators, the relative superiority is:

[0067] (1)

[0068]

[0069] in, For observation point Indicators The measured value, As an indicator The ideal value.

[0070] 2. For interval-type indicators, the relative superiority is:

[0071] (2)

[0072]

[0073] in, and These are the actual measured values ​​of the indicators. The optimal lower and upper limits, for Maximum deviation These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0074] 3. For cost-related indicators, the relative superiority is:

[0075] (3)

[0076] in, These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0077] (4) For benefit-type indicators, the relative superiority is:

[0078] (4)

[0079] in, These are the actual measured values ​​of the indicators. The maximum and minimum values.

[0080] The second part combines set pair theory with radar chart method to present a new comprehensive evaluation method based on the combination of set pair theory and radar chart method (connection number radar chart method).

[0081] 1. Construction of the binary correlation coefficient of indicator values

[0082] After the standardization process for the evaluation indicators was completed in Part 1, the data for each evaluation indicator can be used to construct a binary correlation coefficient of expected variance according to the following definition:

[0083] Definition 1: For a given observation point, let... and Indicators of The expected value and variance of a measured value are called...

[0084] (5)

[0085] As an indicator of The expected variance type of the binary relationship coefficient of a number of measured values ​​is abbreviated as: Where R represents the set of real numbers; when At that time, it was called The standard expected value minus the variance correlation coefficient.

[0086] Definition 2: Correlation coefficient with standard expected variance Then it is called for The modulus, denoted as ;say for The argument, abbreviated as .

[0087] Definition 3: For the general expected value-variance relationship coefficient ,say for The modulus, denoted as ;say for The angle of the argument.

[0088] 2. Contact Number Radar Chart Evaluation Model

[0089] Based on the evaluation indicators of the object to be evaluated, a comprehensive evaluation model based on the radar chart method using the number of connections is presented.

[0090] (1) Based on the content of Part 1, for the object to be evaluated Each observation point has its own set of indicators. The measured data were then standardized.

[0091] (2) Construction of the binary correlation coefficient for the index values. Using the standardized large sample data (measured data), the expected variance binary correlation coefficient for each index at each observation point is calculated using the formulas in Definition 1. Indicates the observation point Indicators The binary relational value.

[0092] Calculate using Definition 3 corresponding module and argument .

[0093] (3) Determine the weight of each indicator Through formula Calculate the central angle corresponding to each indicator on the radar chart; starting from zero degrees, plot the area occupied by each indicator on the radar chart in a counterclockwise direction, in descending order of central angle (e.g., ...). Figure 1 The great circle is divided into seven parts, and the central angle of each part is (the circle is divided into seven parts).

[0094] (4) Draw a radar chart of the correlation coefficients for each indicator, and determine the angle of each binary correlation coefficient. Convert using the following formula:

[0095]

[0096] Using the initial boundary of the area occupied by each indicator as Starting from the edge, draw a circle with a radius of [missing information] in a counter-clockwise direction. (Modulus), central angle is fan-shaped (e.g.) Figure 1 (Middle shaded area).

[0097] (5) Evaluate the merits and demerits of each object to be evaluated. Calculate the total area of ​​each index sector on the radar chart at each observation point. With total perimeter (Hereinafter referred to as area and perimeter)

[0098] (6)

[0099] If the observation point is calculated using the above formula The sector area and perimeter on the contact number-radar chart are respectively... The following rules can be used to evaluate the objects to be evaluated at each observation point:

[0100] (I) If Then it is believed The point to be evaluated is better than , recorded as Conversely, it is written as ;

[0101] (II) If ,but

[0102] ①If Then it is believed The point to be evaluated is better than , recorded as ;

[0103] ②If Then it is believed The point to be evaluated is inferior to , recorded as ;

[0104] ③If Then it is believed Dot and The points to be evaluated are indifferent, denoted as . .

[0105] Based on the implementation steps of the contact number radar chart method given above, it has the following advantages compared with the traditional radar chart method:

[0106] (1) It can realize the evaluation of long-term monitoring data. By calculating the expectation and variance of large sample data, the statistical characteristic values ​​of large sample data are obtained. The expected variance correlation coefficient constructed in this way can reflect the true characteristics of each indicator of the observation point on a long time scale, and the final evaluation result will inevitably be more in line with reality;

[0107] (2) The traditional radar chart method uses the points of each index value to connect to form a polygon (e.g. Figure 2 The traditional radar chart method uses polygons to calculate the area and perimeter of polygons to comprehensively evaluate the objects under assessment at each observation point. However, these polygons occupy a significant amount of irrelevant area, inevitably impacting the evaluation results. In contrast, using the modulus and argument of the expected variance correlation coefficient to plot a sector on the radar chart allows the sector's area and perimeter to accurately represent the statistical values ​​of each indicator, avoiding the errors caused by the polygons occupying irrelevant areas in the traditional radar chart method.

[0108] (3) In the radar chart method of connecting numbers, the sectors that express the statistical values ​​of each indicator fall within the sector area of ​​each indicator after conversion. The sector area and perimeter of the statistical value of the indicator are not related to its specific position in the sector area of ​​each indicator. This avoids the error caused by the inability to clarify the boundaries of adjacent indicators in the traditional radar chart method.

[0109] (4) In the connection number-radar chart method, since the sector area of ​​each index at each observation point is determined by the binary connection number, among which... The value of depends on the scope of the data used in the evaluation of each indicator's measured data, such as... At that time, the range of measured data for the indicators participating in the evaluation was: ,Depend on" "In principle (under normal circumstances)" At this point, the measured data of the indicators used in the evaluation account for approximately 99.74% of the total data. Similarly, when At that time, the range of measured data for the indicators participating in the evaluation was: At this point, the measured data for the indicators used in the evaluation account for approximately 95.44% of the total data. For specific evaluation problems, appropriate methods can be selected based on the required accuracy of the evaluation results. The more stringent the requirements, the higher the value. The larger the value, the greater the proportion of the data used in the evaluation to the total measured data.

[0110] Part Three: An evaluation example is given, comparing the contact number radar chart method with traditional radar chart evaluation methods, which confirms the superiority and feasibility of the method presented in this paper.

[0111] Using power quality assessment as the object of evaluation, four observation points were selected. An assessment was conducted. The power quality assessment indicators included frequency deviation. Waveform distortion rate Voltage deviation Three-phase imbalance Voltage fluctuations and flicker Reliability indicators Service indicators Assume the weights of the seven indicators are: [0.1703, 0.1696, 0.1658, 0.1649, 0.1198, 0.1178, 0.0919].

[0112] The data collected from the four observation points were normalized, and then the binary correlation coefficients of each indicator were constructed. The results are shown in the table below:

[0113] Table 1. Binary correlation coefficients for each indicator at the four observation points.

[0114]

[0115] Based on Table 1, the magnitude and argument of the seven indicators for each observation point were calculated and listed in Tables 2 and 3:

[0116] Table 2 Take At that time, the magnitude (radius) of each indicator at the four observation points.

[0117]

[0118] Table 3. Take At that time, the argument (central angle) of each indicator at the four observation points.

[0119]

[0120] According to equation (6), the following can be calculated respectively: , and At that time, the sector area of ​​the binary correlation coefficient at each observation point on the radar chart , , With perimeter , , Furthermore, using the expected values ​​of each indicator at the observation points as the indicator values, the area of ​​the connected polygon can be calculated using the traditional radar chart method. =[0.3096 0.3134 0.3060 0.2896], the calculation results are summarized in the table below:

[0121] Table 4. Comparison of sector area and perimeter on radar charts under various conditions.

[0122]

[0123] It can be seen from the above table:

[0124] (1) When At that time, the power quality ranking of each observation point was as follows: ;when At that time, the power quality ranking of each observation point was as follows: ;when At that time, observation point and The sector areas of the index values ​​are equal. According to the evaluation rules, the perimeter can only be calculated to determine the ranking of the power quality at the four observation points. The above calculation results show that when When different values ​​are selected, the order of the sector area and perimeter of the calculated index values ​​for each observation point is different. That is, the proportion of data involved in the assessment to the total data will affect the assessment results. Therefore, in actual power quality assessment, it is necessary to select an appropriate value according to the specific assessment requirements. Value, that is, determining the appropriate amount of data to be used in the evaluation;

[0125] (2) Calculation using the traditional radar chart method The results showed that the power quality ranking at the observation points was also... This is related to At this time, the calculation results of the radar chart method are consistent with those of the connection number. According to probability theory, the data involved in the evaluation at this time only accounts for 57.62% of the total data volume, which also shows that there is a deviation in the calculation results of the traditional radar chart method.

[0126] (3) Area calculated by traditional radar chart method The sector area calculated by the radar chart method is almost an order of magnitude larger than that calculated by the connection number radar chart method. This is because the traditional radar chart method adds too much irrelevant area when connecting the index points to form polygons, which will inevitably bring errors to the final comprehensive evaluation results.

[0127] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A comprehensive evaluation method based on a combination of set pair theory and radar chart method, applied to power quality assessment at observation points, characterized in that: Taking power quality assessment as the object of evaluation, the power quality assessment indicators are frequency deviation C1, waveform distortion rate C2, voltage deviation C3, three-phase imbalance C4, voltage fluctuation and flicker C5, reliability index C6, and serviceability index C7. Construct the expected variance type binary connection coefficient for each evaluation index at each observation point, and calculate the modulus and argument; Calculate the central angle of each evaluation indicator in the radar chart based on its weight, and divide the radar chart into corresponding indicator areas based on the central angle of each evaluation indicator. A sector is drawn on the radar chart within the corresponding index area, with the magnitude as the radius and the argument as the central angle. The power quality at each observation point is assessed by comparing the total area and total perimeter of the sectors on each radar chart. The assessment methods include: First, compare the total area of ​​each sector at each observation point; the sector with the larger total area is better than the sector with the smaller total area. When the total area of ​​the observation points is equal, the total perimeter of each sector at each observation point is compared, and the sector with the larger total perimeter is superior to the sector with the smaller total perimeter. When the total perimeter of the observation points is equal, the task is indistinguishable. The binary link coefficient is an expected-variance type binary link coefficient: set up and Observation points Indicators measured value The expected value and variance are then This corresponds to the expected variance type of the binary contact number; The expected variance type binary connection coefficient is divided into two cases: Scenario 1, when hour, Let the standard expected value variance-coefficient be denoted as . for The model, called for The angle of the argument; Scenario 2, when hour, Let Variance be the general expected value-variance correlation coefficient, then for The model, for The angle of the argument; according to Given the given value, select the corresponding case and calculate the corresponding modulus and argument.

2. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 1, characterized in that: Determine the weight of each evaluation indicator Through formula Calculate the corresponding central angle in the radar chart, and starting from zero degrees, plot the index areas on the radar chart in a counterclockwise direction according to the size of the central angle.

3. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 2, characterized in that: Each corner Convert using the following formula: Using the starting boundary of each indicator region as Starting from the edge, draw a circle with a radius of modulo 1 in a counterclockwise direction. The central angle is A sector.

4. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 1, characterized in that: Before constructing the expected variance type binary correlation coefficient, the evaluation indicators are standardized: the evaluation indicators are divided into four types: interval indicators, fixed indicators, benefit indicators and cost indicators, and a standardized model of the evaluation indicators is given.

5. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 4, characterized in that: The normalized model for the fixed index is as follows: For fixed indicators, the relative superiority is: in, For observation point Indicators The measured value, As an indicator The ideal value.

6. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 4, characterized in that: The normalized model for the interval-type indicator is as follows: For interval-type indicators, the relative superiority is: in, For observation point Indicators The measured value, and These are the actual measured values ​​of the indicators. The optimal lower and upper limits, for Maximum deviation These are the actual measured values ​​of the indicators. The maximum and minimum values.

7. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 4, characterized in that: The standardized model for the cost-type indicators is as follows: For cost-related indicators, the relative superiority is: in, For observation point Indicators The measured value, These are the actual measured values ​​of the indicators. The maximum and minimum values.

8. The comprehensive evaluation method based on the combination of set pair theory and radar chart method according to claim 4, characterized in that: The standardized model for the aforementioned benefit-type indicators is as follows: For benefit-oriented indicators, their relative superiority is: in, For observation point Indicators The measured value, These are the actual measured values ​​of the indicators. The maximum and minimum values.

Images