A large-scale heterogeneous multi-attribute equipment scheme clustering decision method and medium

Abstract

The application relates to a large-scale heterogeneous multi-attribute equipment scheme clustering decision method and medium, which comprises the following steps: converting heterogeneous attribute data in a digital twin decision information system into binary connection numbers; determining attribute weights by adopting a MEREC method; constructing attribute irreplaceable values based on the attribute weights and the binary connection numbers; generating a scheme similar network according to the attribute irreplaceable values; performing network clustering operation on the scheme similar network to output a scheme clustering result; and calculating the comprehensive target heart distance of the schemes in the cluster based on the scheme clustering result by adopting a grey target theory, and generating a final scheme score by fusing the scheme global value and the comprehensive target heart distance. The method solves the problems of heterogeneous information fusion distortion, attribute irreplaceable mechanism loss and low large-scale scheme clustering efficiency in a digital twin decision scene by constructing a 'heterogeneous data unification-attribute irreplaceable mechanism modeling-network clustering-clustered decision' mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of complex equipment decision optimization, in particular to a large-scale heterogeneous multi-attribute equipment scheme clustering decision method and medium. BACKGROUND

[0002] Under the background of digital twin technology being widely applied in high-end equipment demonstration, the decision information system faces massive heterogeneous (such as precise numbers, interval numbers, sequence numbers, language variables, various fuzzy numbers and language term sets, etc.) and multi-scale evaluation data, and meanwhile, a large number of alternative schemes are involved in the equipment demonstration process, forming a typical large-scale heterogeneous multi-attribute decision problem. The existing researches have significant limitations in dealing with such problems: on the one hand, the fusion methods for heterogeneous information are mainly divided into two categories based on conversion (such as conversion into fuzzy numbers, interval numbers, etc.) and distance (such as distance from ideal solution), the former is easy to cause information loss in the conversion process, and the latter over-reliance on the accuracy of reference point selection and difficult to ensure the quality of homogeneous preference values, leading to decision bias; on the other hand, in the face of large-scale scheme set, the traditional decision-making method has high computational complexity and low efficiency, and the existing simplification techniques (such as attribute reduction, scheme clustering, and efficient sorting algorithm) mainly focus on a single dimension (attribute heterogeneity or scheme scale), lacking of integrated methods that can effectively handle heterogeneous information fusion and large-scale scheme clustering decision, which seriously restricts the effectiveness and practicality of decision-making.

[0003] Specifically, the existing technology has three key defects: first, the information homogenization process mainly focuses on the evaluation information expressed by the decision maker, and lacks effective unified transformation methods for sensor measurement information (such as feature parameters) collected by machines, and it is difficult to fully reflect the inherent fuzziness and uncertainty in the comparison between schemes; second, the attribute measure method is too simple (often relying on a single distance or similarity), and fails to consider multiple dimensions of attribute characteristics such as distance, similarity, correlation and entropy, especially ignoring the attribute non-compensable mechanism (i.e. the short board of one attribute cannot be compensated by the advantage of other attributes) commonly existing in high-end equipment, leading to the loss of attribute characteristics and evaluation distortion between schemes; third, there is insufficient research on clustering decision for large-scale scheme characteristics, lacking of quantitative clustering standard function, making it difficult to design a unified standard to realize scientific classification of schemes and effectively coordinate the unity of global decision and local decision, and unable to fully utilize the concept of "uniform evaluation of similar schemes" to improve decision efficiency.

[0004] Therefore, it is urgent to develop a new large-scale heterogeneous multi-attribute high-end equipment scheme clustering decision method for digital twin scenarios to overcome the above defects. SUMMARY

[0005] The application provides a large-scale heterogeneous multi-attribute equipment scheme clustering decision method and a medium, which aims to solve the problems of heterogeneous information fusion distortion, attribute non-compensable mechanism loss and low efficiency of large-scale scheme clustering in a digital twin decision-making scenario.

[0006] To achieve the above-mentioned purpose, the first aspect of the application provides a large-scale heterogeneous multi-attribute equipment scheme clustering decision method, comprising the following steps:

[0007] Uniformly converting heterogeneous attribute data in a digital twin decision-making information system into a binary connection number;

[0008] Determining attribute weights by using a MEREC method, and constructing attribute non-compensable values based on the attribute weights and the binary connection number;

[0009] Generating a scheme similarity network according to the attribute non-compensable values;

[0010] Performing a network clustering operation on the scheme similarity network, and outputting a scheme clustering result;

[0011] Based on the scheme clustering result, calculating a comprehensive target center distance of schemes in a cluster by using a grey target theory, and generating a final scheme score by fusing a global value of the scheme and the comprehensive target center distance.

[0012] Further, the heterogeneous attribute data includes sensor measurement information, and the method for converting the sensor measurement information into a binary connection number comprises:

[0013] Calculating the standardized distance of a candidate scheme and each scheme in a target database under the same attribute;

[0014] According to the standardized distance, calculating the positive evidence support degree and the negative evidence support degree of the candidate scheme, respectively, wherein the positive evidence support degree represents the degree that the candidate scheme is better than other schemes, and the negative evidence support degree represents the degree that the candidate scheme is worse than other schemes;

[0015] Constructing a confidence interval based on the positive evidence support degree and the negative evidence support degree, wherein the confidence interval contains the trust degree and the non-negative degree of the attribute of the scheme;

[0016] Mapping the confidence interval to a certain-uncertain space, extracting the certainty component and the uncertainty component, and forming a binary connection number.

[0017] Further, the heterogeneous attribute data includes expert evaluation information; the expert evaluation information is converted into a binary connection number, and the following operations are performed according to the information type:

[0018] For a language variable, a qualitative language is quantified by using a language scale function to generate a quantitative value, and the hesitancy degree is calculated, and the binary connection number is generated by combining the quantitative value and the hesitancy degree;

[0019] For hesitant fuzzy numbers, the geometric mean hesitancy degree is calculated, and the binary connection number is generated by combining the membership degree score and the geometric mean hesitancy degree;

[0020] For a set of probabilistic linguistic terms, the quantification value of the linguistic term is calculated by a score function, and the binary connection number is generated by combining the uncertainty represented by the deviation function.

[0021] Further, the method for determining attribute weights using the MEREC method includes:

[0022] Extract the certainty component and the uncertainty component in the binary connection number;

[0023] Based on the certainty component, calculate the distance measure, similarity measure, correlation measure, and entropy measure;

[0024] By weighting and fusing the distance measure, similarity measure, correlation measure, and entropy measure by the attribute weight, the attribute non-redeemable value representing the attribute consistency difference between schemes is generated.

[0025] Further, the method for determining attribute weights using the MEREC method includes:

[0026] For benefit-type attributes, normalize the attribute values by the maximum value of each attribute; for cost-type attributes, normalize the attribute values by taking the reciprocal of the minimum value of each attribute;

[0027] Take the natural logarithm of the normalized attribute values in the benefit-type attributes or cost-type attributes and sum them up to obtain the total performance value of each scheme;

[0028] After removing each attribute in turn, recalculate the total performance value of each scheme after removal;

[0029] For each attribute, accumulate the absolute deviation of the original total performance value and the total performance value after removing the attribute;

[0030] The absolute deviation of each attribute and the proportion of the sum of all attribute absolute deviations are taken as the attribute weight value of the attribute.

[0031] Further, the network clustering operation includes:

[0032] Using a game combination optimization strategy, fuse multiple node similarity calculation methods to generate a comprehensive similarity matrix;

[0033] Perform Markov clustering on the comprehensive similarity matrix, and sequentially perform network expansion, weight inflation, and probability pruning;

[0034] Re-examine the neighborhood link relationship of the boundary nodes in the clustering result, and adjust the cluster group to which the nodes belong;

[0035] Merge the cluster groups with too small size, and output the final scheme clustering result.

[0036] Further, the method for calculating the comprehensive target distance of the intra-cluster scheme comprises:

[0037] Perform reward-punishment transformation on the attribute values of the intra-cluster scheme, wherein the benefit type attribute adopts an upper limit effect measure, and is standardized by taking the maximum value of the attribute as a reference; and the cost type attribute adopts a lower limit effect measure, and is standardized by taking the minimum value of the attribute as a reference.

[0038] Based on the standardized attribute values, determine the positive target center and the negative target center of the intra-cluster scheme, wherein the positive target center is composed of optimal values of each attribute, and the negative target center is composed of worst values of each attribute.

[0039] Calculate the positive target distance of each scheme from the positive target center and the negative target distance of each scheme from the negative target center.

[0040] Take the connecting line of the positive and negative target centers as a reference axis, and calculate the projection value of the positive target distance of each scheme on the axis as the comprehensive target distance.

[0041] Further, the scheme global value is obtained by the following method:

[0042] Determine a positive ideal solution based on the attribute values of all alternative schemes, wherein each attribute takes an optimal value, and for a benefit type attribute, takes a maximum value, and for a cost type attribute, takes a minimum value.

[0043] Determine a negative ideal solution based on the attribute values of all alternative schemes, wherein each attribute takes a worst value, and for a benefit type attribute, takes a minimum value, and for a cost type attribute, takes a maximum value.

[0044] For each scheme, calculate the distance of the scheme from the positive ideal solution, which is obtained by a weighted sum of the difference values of the attribute values, and the weight is the attribute weight.

[0045] For each scheme, calculate the distance of the scheme from the negative ideal solution, which is obtained by a weighted sum of the difference values of the attribute values, and the weight is the attribute weight.

[0046] Divide the distance of the scheme from the negative ideal solution by the sum of the distance of the scheme from the positive ideal solution and the distance of the scheme from the negative ideal solution to obtain the scheme global value.

[0047] Further, the final scheme score is generated by weighted fusion of the scheme global value and the comprehensive target distance, and the weight coefficient is dynamically determined by the size proportion of the scheme subgroup.

[0048] To achieve the above object, the third aspect of the present application provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program performs the steps of the large-scale heterogeneous multi-attribute equipment scheme clustering decision method when run by a processor.

[0049] The present application has the following advantages:

[0050] Compared with the prior art, the large-scale heterogeneous multi-attribute equipment scheme clustering decision method and medium provided by the present application solve the above problems by constructing a whole-process method system of "heterogeneous data unification-attribute non-compensable mechanism modeling-network clustering-in-cluster decision": first, for the distortion of heterogeneous information fusion, a unified transformation mechanism based on D-U space mapping is proposed, which converts sensor measurement information (exact numbers, interval numbers, sequence numbers) and expert evaluation information (language variables, hesitant fuzzy numbers, probability language term sets, etc.) into binary connection number form. By separating the certain component and the uncertain component (such as mapping the confidence interval into a certain boundary point and an uncertain interval), the original data fuzziness and structural characteristics are preserved while achieving homogeneous information without loss; second, for the lack of attribute non-compensable mechanism, a multi-dimensional measurement model is constructed, the attribute weights are calculated by MEREC method, and the attribute non-compensable values between schemes are quantified by combining distance measure, similarity measure, correlation measure and entropy measure. The global scheme value is generated by comparing the multi-dimensional feature gap between the scheme and the positive and negative ideal solutions, which accurately describes the nonlinear dependence and short board effect between attributes; finally, for the low efficiency of large-scale clustering, a scheme similarity network is constructed based on the attribute non-compensable value, a game combination optimization strategy is used to integrate multiple node similarities, and an improved MCL algorithm is used to realize scheme clustering (including boundary rechecking and small cluster merging optimization), which significantly reduces the decision scale. The comprehensive target distance is calculated in the cluster by using the grey target theory, and finally the ranking is generated by fusing the global value of the scheme and the variable weight score of the local target distance, which reduces the computational complexity while ensuring the decision accuracy, and realizes efficient and accurate decision. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows.

[0052] Figure 1 is a large-scale heterogeneous multi-attribute equipment scheme clustering decision method framework disclosed by the embodiments of the present application.

[0053] Figure 2 is a digital twin five-dimensional model structure diagram disclosed by the embodiments of the present application.

[0054] Figure 3 is a D-U space coordinate diagram disclosed by the embodiments of the present application.

[0055] Figure 4 is a mapping diagram of sensor measurement confidence interval on D-U space disclosed by the embodiment of the present application.

[0056] Figure 5 is a mapping diagram of expert evaluation information D-H space disclosed by the embodiment of the present application. Figure 5 (a) is a D-U space mapping diagram of sensor measurement information, Figure 5 (b) is a D-U space mapping diagram of expert evaluation language variable, Figure 5 (c) is a D-U space mapping diagram of probability language term set.

[0057] Figure 6 is a high-end equipment scheme optimization problem analysis diagram under the influence of attribute non-replacement mechanism disclosed by the embodiment of the present application.

[0058] Figure 7 is a project clustering network example diagram disclosed by the embodiment of the present application.

[0059] Figure 8 is a comprehensive target distance diagram disclosed by the embodiment of the present application.

[0060] Figure 9 is a heat map of attribute non-replacement value between two computing schemes disclosed by the embodiment of the present application.

[0061] Figure 10 is a scheme clustering decision network diagram disclosed by the embodiment of the present application.

[0062] Figure 11 is a network diagram after clustering disclosed by the embodiment of the present application. DETAILED DESCRIPTION

[0063] In order to make the person skilled in the art better understand the present application scheme, the technical scheme in the embodiment of the present application will be described clearly and completely below in conjunction with the drawings in the embodiment of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by the person skilled in the art without creative labor should belong to the scope of protection of the present application.

[0064] According to the embodiment of the present application, it should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the following manufacturing method, in some cases, the steps shown or described can be executed in an order different from here.

[0065] In the specific implementation of this invention, it is crucial to understand the fundamental concepts of digital twins and the basic principles of set pair analysis (SPA) and deterministic-uncertain space (DU space), which are described in detail below:

[0066] Digital twins refer to the creation of virtual models of physical entities in a virtual space by fully utilizing data such as physical models, sensor updates, and operational history, integrating multi-disciplinary, multi-physical quantity, multi-scale, and multi-probabilistic simulation processes, and simulating the behavior of these physical entities in the real environment. As a virtual mirror image of the physical entity, it reflects the entire lifecycle of the corresponding physical entity product, adding or expanding new capabilities for the physical entity. In the high-end equipment demonstration scenario addressed by this invention, digital twins offer advantages of high efficiency and rapid iteration. The widely adopted five-dimensional digital twin model (physical entity, virtual entity, connectivity, twin data, and services) includes:

[0067] (1)

[0068] in, It is a multidimensional evaluation vector. Represents physical entities, Represents a virtual entity. Indicates service, Representing twin data, This represents the connections between the various components. The structure of the five-dimensional digital twin model is as follows: Figure 2 As shown, in the process of using digital twins to demonstrate equipment schemes, the decision information data structure is often multi-layered, involving relationships at different scales and levels, which traditional information systems find difficult to capture.

[0069] When digital twins are used for equipment scheme demonstration, their decision information data structure has multi-level and multi-scale correlation characteristics. This invention is based on the following formal definition:

[0070] Definition 1: [This is a term used in Chinese, and the translation reflects that.] As a digital twin decision information system, in which It is a non-empty finite set of objects (i.e., a set of alternative solutions). It is a non-empty finite set of attributes. It is a non-empty finite-scale set.

[0071] Any alternative In attributes Different information scales Different evaluation values ​​exist. Therefore, a digital twin decision information system can also be represented as: .

[0072] There are two kinds of data in digital twin decision information system, which are sensor measurement information and expert decision information. The characteristic parameter measurement value of the sensor is the characteristic layer information, and the experience knowledge of the expert is the decision layer information. The data given by the sensor is the measurement value of a certain characteristic attribute of the scheme, which represents the extension property of the target scheme. Therefore, the measurement value of the sensor needs to be converted into the degree of support or opposition to the scheme, and the expert decision information needs to be converted into the same scale.

[0073] Definition 2: Hypothesis is a digital twin decision information system, where is a surjection ( is an attribute whose value range is in the first scale. There is a surjection such that , where is called the information scale transformation function from scale to scale under attribute .

[0074] Definition 3: Set pair analysis (SPA) is a theory that studies the interaction between certainty and uncertainty in nature and human society. Set pair is a pair of subsystems composed of two related sets. SPA analyzes the identity (a), difference (b), and opposition (c) in the set pair system, which opens up new ideas for solving random uncertainty problems and fuzzy uncertainty problems. Connection number is the main mathematical tool of SPA, and its general expression is:

[0075]

[0076] where is used to handle the uncertainty of the characteristics, , and are any non-negative real numbers. For a certain attribute, they are mathematical measures of the degree of identity, difference, and opposition, respectively. is the opposition coefficient, which represents the opposition between and . is the difference coefficient, which takes the value of the uncertainty between -1 and 1 through , representing the uncertainty of the attribute. According to the different uncertainties in the problem, can be extended to the form of . In the case of not considering the opposition of the set pair system, the connection number can be simplified to the form of . This form is suitable for handling the fuzziness and uncertainty in decision information.

[0077] Deterministic-uncertain space is a space containing uncertainty, as defined by SPA, characterized by its composition. 3D space At least one dimension of the coordinate axis describes the measure of uncertainty. Figure 3 The composite quantity containing uncertain components in the two-dimensional DU space shown For example, in the figure Axis Description Deterministic components, Axis Description The uncertainty component.

[0078] The essence of the theoretical basis of the MCL algorithm is to calculate the process of random walks in graph G by Markov chains, which mainly consists of three stages: expansion, dilation and pruning.

[0079] Definition 4: Define a network diagram of a community structure as follows This indicates that the network diagram Adding a ring to the adjacency matrix Standardization enables The range of edge set elements is ,and .use The state transition matrix is ​​represented by the following formula:

[0080] (2)

[0081] in, Let be the state transition matrix. Middle node To the node The transition probabilities (the matrix after the expansion operation). Let be the adjacency matrix of the network. It is the identity matrix. For spatial location index, This represents the maximum number of rows. Adding a ring to a matrix means adding an edge to each node, which in turn adds 1 to each element on the diagonal. Normalizing the matrix... Processing, among which Represents the identity matrix; Standardization factors are listed.

[0082] The expansion operation is the process of multiplying two matrices by themselves, expressed in the following formula:

[0083] (3)

[0084] in, This represents the expanded transition matrix. is a binary function; , is a feature vector / observation; is a normalized transition matrix, is an expansion parameter.

[0085] When the expansion value is larger, the random walk is more likely to leave the cluster it is in and enter the remaining clusters. Therefore, as the expansion value increases, the connection between the same cluster is weakened.

[0086] The inflation operation is to multiply the elements in the corresponding positions of the two matrices and normalize the columns of the matrix, that is, to multiply the elements in the corresponding positions of the two matrices and then normalize the columns of the matrix. The formula is:

[0087] (4)

[0088] (5)

[0089] wherein, is a state transition matrix is the transition probability of the node to the node after the expansion operation (the matrix after the expansion operation); is an inflation parameter; is the total number of nodes in the network; represents the probability of the node transitioning to the community after the inflation operation; is the transition probability matrix after inflation; is the new state transition matrix after the inflation operation, which is used for the next pruning.

[0090] The accurate pruning is to retain values with larger probabilities, and the probability values of the remaining nodes are zero:

[0091] (6)

[0092] wherein, represents the matrix after pruning, is the matrix after the inflation operation, is a pruning parameter, and the elements in the matrix are arranged in descending order in each column. All elements after the th element are set to 0 to obtain the matrix . The expansion, inflation and pruning operations are iteratively performed until convergence, and the final matrix can be mapped to obtain the clustering (community) structure of the network. The algorithm provides core method support for the network clustering operation in step S400 of the present application.

[0093] As shown in the figure, the application provides a large-scale heterogeneous multi-attribute equipment scheme clustering decision method, comprising the following steps: Figure 1

[0094] Step S100, uniformly transforming heterogeneous attribute data in a digital twin decision information system into binary connection numbers;

[0095] Step S200, determining attribute weights by using the MEREC method, and constructing attribute non-compensable values based on the attribute weights and the binary connection numbers;

[0096] Step S300, generating a scheme similarity network according to the attribute non-compensable values;

[0097] Step S400, performing a network clustering operation on the scheme similarity network, and outputting a scheme clustering result;

[0098] Step S500, calculating a comprehensive target center distance of schemes in a cluster based on the scheme clustering result by using the grey target theory, and generating a final scheme score by fusing a global value of the scheme and the comprehensive target center distance.

[0099] In one specific embodiment of the application, as described in the above step S100, heterogeneous attribute data in a digital twin decision information system mainly comes from sensor measurement information and expert evaluation information, and needs to be uniformly transformed into binary connection numbers . The specific processing mode is different according to the data type. Generally, there are three types of high-end equipment feature data in a digital twin scene, namely, accurate numbers, interval numbers and sequence numbers.

[0100] The attribute of the accurate number type is represented by a single real value to represent the attribute of the equipment. The interval number attribute is used to describe the range type attribute of the equipment, which is an effective description means for handling complex and uncertain information. The sequence number represents the sequence data of the equipment changing with the characteristics in the time or space dimension, which is usually a continuous or ordered numerical sequence.

[0101] Let the mixed attribute characteristic parameter vector of the th high-end equipment scheme be , the mixed characteristic parameter vector of the th target scheme in the equipment target parameter database be , the distance between the th unknown target characteristic data and the characteristic parameter in the target database be , The Euclidean distance is used for normalization processing: ​

[0102] (7)

[0103] wherein the symbol and denote the scheme number, which is uniform in value, denotes the characteristic attribute number condition; is the equipment scheme and the target scheme is the original distance (such as the Euclidean distance) under the attribute ; is the characteristic parameter value of the high-end equipment scheme under the attribute ; is the characteristic value of the target scheme (reference scheme) under the attribute ; is the distance type identifier; is the standardized distance value.

[0104] Considering the equipment characteristic criterion and a standardized distance vector , assuming that the equipment characteristic attributes are all benefit types, the equipment scheme set , for any one scheme in , the positive evidence support and the negative evidence support thereof are defined as follows:

[0105] (8)

[0106] (9)

[0107] wherein, is the positive evidence support; is the negative evidence support; is the alternative scheme, the index of which is ; is the index variable of other schemes; is the normalized distance value of other schemes under the attribute and the scale ; is the normalized distance value of the current scheme under the attribute and the scale ;

[0108] Let , , if and are not both 0, then there exists:

[0109] (10)

[0110] where, the positive support, measuring the degree of advantage over other alternatives on attribute the negative support, measuring the degree of disadvantage over other alternatives on attribute is the maximum value of the positive support on attribute is the minimum value of the negative support on attribute

[0111] The confidence interval of selecting scheme on attribute is:

[0112] (11)

[0113] where, represents the degree of trust in represents the degree of non-negation in

[0114] The mapping of the confidence interval of selecting scheme on attribute in the D-U space, the confidence interval is an interval number, denoted as , where, , and when , i.e. the point is a real number. With the help of connection numbers, interval numbers are mapped on the D-U space:

[0115] (12)

[0116] where, is the overall confidence state, composed of the real part (certainty) and the imaginary part (uncertainty), the real part represents the degree of trust in the scheme; the imaginary part represents the comprehensive uncertainty of the scheme, is the uncertainty coefficient, taking the value range of , for the given , the connection number form is obtained by processing:

[0117] ​​​​​​​​​​​ (13)

[0118] where, is the confidence connection number.

[0119] The confidence interval of sensor measurement data is mapped onto the two-dimensional D-U space as shown in Figure 4 In Figure 4 , the boundary points of the confidence interval number and are determined, and the points between and are uncertain, so that the deterministic part can be represented as , and the uncertain part is represented as .

[0120] Expert evaluation decision information is divided into three types of language variables, fuzzy values, and language terms. Language variables refer to evaluation information represented by specific numerical values or symbols. Language variables are used for equipment attributes that can be clearly quantified. Fuzzy values are used to describe evaluations of equipment attributes with certain uncertainty or complexity. Language variables are a qualitative evaluation method suitable for equipment attributes that are difficult to accurately quantify and are expressed in natural language.

[0121] (1) Representation of language variables in D-U space

[0122] Expert evaluation decision information precise language variable type attributes are usually represented by a single language symbol , which usually represents the abbreviation of a certain type of equipment attribute.

[0123] is a set of an odd number of language variables, which quantifies the qualitative indicators for evaluating language variables, the language scale function of

[0124] (14)

[0125] It can be understood that the function of this formula is to map the discrete semantic items in the language variable to a deterministic membership degree (such as "support degree").

[0126] where, is the language item, is the offset parameter, (representing the offset degree of the semantic item), (normalized variable); is the total number of language items; is the deterministic membership degree.

[0127] Then language variables The degree of hesitation is:

[0128] (15)

[0129] in, The degree of hesitation.

[0130] There are language variables ,make , The obtained connection number form According to this method, linguistic variables are converted into binary relational expressions.

[0131] (2) Representation of hesitant fuzzy numbers in DU space

[0132] The fuzzy numbers in expert evaluation decision-making information include types such as triangular fuzzy numbers, trapezoidal fuzzy numbers, discrete fuzzy numbers, intuitive fuzzy numbers, and hesitant fuzzy numbers, meaning that a certain degree of fuzziness can be included when providing attribute evaluations of equipment.

[0133] For fuzzy number attributes, taking hesitant fuzzy numbers as an example, hesitant fuzzy elements are used to represent the degree of hesitation among experts when evaluating the attribute. The mathematical expression of hesitant fuzzy elements is as follows: ,in It is a set of values ​​in [0,1], called For hesitant and ambiguous elements.

[0134] To fully account for the degree of hesitation among decision-makers, this embodiment considers both the differences in membership between elements in the hesitant fuzzy element and the different numbers of elements, and defines a geometric mean degree of hesitation to characterize the uncertain information of hesitant fuzziness.

[0135] Given that the geometric mean hesitation of an HFE is defined as:

[0136] (16)

[0137] in, For hesitant and blurred elements, The number of elements. yes The scoring function for each element is as follows:

[0138] (17)

[0139] Among them, if ,but Given a hesitant fuzzy number. ,make , The obtained connection number form .

[0140] (3) Representation of the probabilistic language term set in the DU space

[0141] The linguistic terminology includes various types such as hesitant and ambiguous terminology sets, two-layer terminology sets, probabilistic terminology sets, and multi-granularity terminology sets. These are used to help experts evaluate complex and uncertain attributes, providing a rich set of means for the multi-layered and ambiguous expression of equipment attributes. This embodiment uses a probabilistic terminology set as an example, employing probabilistic terminology to represent the linguistic preferences of experts when evaluating attributes.

[0142] Let LTS be PLTS is represented as:

[0143] (18)

[0144] in, Basic language set, For indexing language terms, It is a linguistic term. yes The corresponding probability, express Number, This is a probability normalization constraint.

[0145] set up For any PLTS language term, the language term The small mark is The score function and its inverse function for a single language are defined as follows:

[0146] (19)

[0147]

[0148] in, for In the language collection Subscript in; Probability-weighted semantic scores; Indicates the continuous semantic strength of language terms after standardization; A semantic description in numerical form. Indicates continuous value input; Based on numerical values Reconstructing language terminology ( (for integer order); This is a set of probabilistic language terms.

[0149] Conversion logic: Forward conversion Language item Rounding mapping Standard value ;

[0150] Reverse conversion : Numerical value Reconstruction function Language item

[0151] The score function of is defined as:

[0152] (20)

[0153] where, is the PLTS overall score (for D-U space mapping); is the discrete value after rounding the single language term; is the probability weighting of the discrete semantic score.

[0154] Let be any one PLTS, the small mark of language term is , The bias function of is defined as:

[0155] (21)

[0156] Let there be a probability language set , let , The resulting connection number form is as follows:

[0157] (22)

[0158] The mapping of language variables , hesitant fuzzy numbers and probability language sets on the two-dimensional D-U space is shown in Figure 5 According to the content of the figure and the correlation analysis of the formula, (a), (b), (c) in the figure respectively represent the D-U space mapping relationship of the following three types of data, which correspond to the following: (a) is the D-U space mapping of sensor measurement information; (b) is the D-U space mapping of expert evaluation language variables; (c) is the D-U space mapping of the probability language term set (PLTS).

[0159] In one specific embodiment of the present application, as described in step S200 above, in a complex equipment or system, the short board or defect of a certain specific attribute cannot be compensated by the advantages of the system in other attributes. This is manifested by the relative independence of the performance values between different attributes, each attribute has its indispensable role in realizing the overall function and cannot be simply replaced or compensated by the performance of other attributes. The interrelation between attributes cannot be considered as linear or additive, but as a complex dependence and relative independence.

[0160] As shown in Figure 6 , the traditional distance measure-based method can evaluate the overall ability of the high-end equipment scheme, and select the scheme closest to the ideal solution. The data statistics correlation measure-based method can evaluate the similarity of the ability gap in different attributes, and select the scheme similar to the ideal solution. The information entropy-based measure quantifies the information amount and uncertainty of different attributes in the scheme, and provides a reference for evaluating the stability of the overall scheme. A high entropy value can represent the diversity and complexity of attribute performance, reflecting the adaptability of the scheme in different environments. The correlation analysis-based measure represents the correlation strength and direction between two or more attributes, and measures their relevance.

[0161] Considering the influence of the non-compensable attribute mechanism, this embodiment comprehensively considers distance measure, similarity measure, correlation measure and entropy measure, and considers that a better high-end equipment scheme should be a scheme that is very "similar" to the ideal scheme. The understanding of "similar" is that the distance between the development scheme and the ideal scheme is small, and the change trend of the development scheme in each attribute is similar to that of the ideal scheme. The correlation measure is a measure used to measure the relationship strength and direction between variables, which helps to identify the interaction and dependence between different attributes in high-end equipment evaluation and data analysis, and further evaluate the effectiveness of the ability arrangement. At the same time, it satisfies the characteristics of a certain complexity, which characterizes the diversity and complexity of attributes, and reflects the adaptability of the scheme in different environments.

[0162] If the D-U space has a parameter , the quantitative value represents the proportion of certainty, which characterizes the numerical size of the D-U space, and is calculated as follows:

[0163] (23)

[0164] wherein, is a real quantitative value, is a certainty value, is an uncertainty value, .

[0165] If the D-U space has a parameter , the generalized standardized D-U space distance is calculated as follows:

[0166] (24)

[0167] where, is the generalized normalized distance; is the quantized value of scheme A on the th attribute; is the quantized value of scheme B on the th attribute; in particular, is the number of attributes, , is the scheme A, is the scheme B, is the attribute weight, is the distance norm parameter, if , the D-U space distance degenerates to the Hamming distance; if , it degenerates to the Euclidean distance.

[0168] Referring to the cosine similarity, if the D-U space has a parameter , the similarity measure of the D-U space , the calculation formula is as follows:

[0169] (25)

[0170] If the D-U space has a parameter , the correlation measure of the D-U space element , the calculation is as follows:

[0171] (26)

[0172] where, , is the mean, , .

[0173] Referring to the definition of relative entropy measure, if the D-U space has a parameter , the entropy measure of the D-U space , which is used to measure the difference between two probability distributions and describes the “information loss” from one distribution to another, is calculated as follows:

[0174] (27)

[0175] The non-recoverable value of the attribute in the D-U space , which is used to measure the difference in attribute consistency between schemes, is calculated as follows:

[0176] (28)

[0177] wherein, is the non-compensable value; are the 4 core evaluation dimensions (i.e. distance measure, similarity measure, correlation measure and entropy measure) of the corresponding equipment scheme.

[0178] The optimal value and the worst value of the attribute in all schemes respectively constitute the positive ideal solution and the negative ideal solution as the positive and negative reference points.

[0179] (29)

[0180] (30)

[0181] wherein, is the positive ideal solution, is the attribute quantization value; is the attribute index; is the attribute set; is the negative ideal solution.

[0182] Then for the equipment selection scheme containing attributes , the scheme attribute feature gap of each scheme with the positive and negative ideal solutions is defined respectively as:

[0183] (31)

[0184] wherein, is the positive distance; is the negative distance; is the single-dimensional distance function; is the positive ideal solution parameter; is the negative ideal solution parameter; is the scheme to be evaluated in the corresponding dimension quantization value.

[0185] Further calculate the scheme global value is:

[0186] (32)

[0187] wherein, is the scheme global value; is the scheme identification. Score as the quantization value of the scheme for the next step of evaluation decision.

[0188] The MEREC weight determination method based on D-U space is as follows:

[0189] This method obtains attribute weight information by measuring the effect of removing each attribute on the overall performance of the candidate solutions. In this method, the greater the impact of deleting an attribute on the overall performance of the candidate solutions, the greater the weight it should be assigned. Based on this idea, this embodiment designs a MEREC weight determination method based on information entropy, with the following specific steps:

[0190] Step S201: Standardize the initial evaluation matrix. If the attribute is benefit-type, then:

[0191] (33)

[0192] If the attribute is of type cost, then:

[0193] (34)

[0194] in, This is the original evaluation value; Standardized evaluation values; For the first of all schemes The maximum value (benefit type) or minimum value (cost type) of an attribute; For the first of all schemes The minimum (benefit-oriented) or maximum (cost-oriented) value of an attribute; The number of alternative options; To evaluate the number of attributes.

[0195] Step S202: Calculate the overall performance of each alternative solution. :

[0196] (35)

[0197] in, For the first The overall performance index of the scheme is as follows: the larger the value, the better the scheme. This is a logarithmic transformation function.

[0198] Step S203: Calculate the performance of the solution after removing each attribute. :

[0199] (36)

[0200] in, Indicates the first The alternative is to remove the first option. Overall performance after considering all attributes; This represents the total number of remaining attributes after removing the attributes. Index for the remaining attributes.

[0201] Step S204: Calculate the sum of absolute deviations for each attribute:

[0202] (37)

[0203] in, Indicates the first The sum of absolute deviations of each attribute; This is a performance deviation. This is the scheme index.

[0204] Step S205: Obtain the objective weight of the attribute:

[0205] (38)

[0206] in, Indicates the first The objective weights of each attribute satisfy and .

[0207] In a specific embodiment of the present invention, as described in step S300 above, this step is based on the attribute non-compensable value calculated in step S200. (To represent the differences in attribute consistency among schemes), a scheme similarity network is constructed. This network treats candidate schemes as nodes and similarity relationships between schemes as edges, providing the basic structure for subsequent clustering decisions. The specific implementation process is as follows:

[0208] In step S100, the heterogeneous attribute data (including sensor measurement information and expert evaluation information) has been uniformly converted into binary relational numbers. ,in, For deterministic components, This is an uncertain component. In step S200:

[0209] The MEREC method was used to determine the weights of each attribute. Based on weights and binary relationship coefficients, an attribute non-compensatory value is constructed. Attributes that cannot be compensated The plan was quantified. and The smaller the value, the higher the similarity between the schemes.

[0210] Based on the irreplaceable value of the attribute The process of constructing a network similar to the proposed solution is as follows:

[0211] Step S301: Traverse all solution pairs Calculate its attribute non-compensable value This value directly reflects the differences between the options: The smaller the value, the more similar the solutions.

[0212] Step S302: To control network density, set a threshold. This threshold is based on all schemes. The distribution of values ​​is determined, and the 75th percentile is usually selected (i.e., only the 25% of connections with the highest similarity are retained). When At that time, in the plan and Add an edge between them; otherwise, there is no connection.

[0213] Step S303: Use all alternative solutions as a node set. ,satisfy The scheme is for edge set This forms an undirected weighted graph. The edge weight is... The lower the weight, the stronger the similarity. To visually illustrate the network structure, Figure 7 The threshold is indicated. Similar networks to schemes 1, 4, and 5. For example, scheme 3 is similar to schemes 1, 4, and 5. Therefore, there are connecting edges; Scheme 2 and Scheme 3 Therefore, there is no connection.

[0214] In a specific embodiment of the present invention, as described in step S400 above, for large-scale solution decision-making problems, the present invention proposes a node similarity network clustering method based on game-theoretic combinatorial optimization. This method improves clustering accuracy by fusing multiple similarity metrics to avoid information loss from single methods. The core idea is to treat different similarity calculations as "game participants" and achieve a consistent optimal similarity assessment through collaborative optimization. The specific network clustering operation is as follows:

[0215] Assuming to adopt Various node similarity calculation methods (such as cosine similarity, Jaccard coefficient, etc.) are used to perform pairwise similarity calculations on network nodes, generating a basic similarity set: A comprehensive similarity vector is constructed through linear combination:

[0216] (39)

[0217] In the formula, This represents the total number of node similarity algorithms used (such as cosine similarity, Jaccard coefficient, etc.). For algorithm indexing, Based on the similarity vector; To synthesize the similarity vector, The coefficients of the linear combination to be optimized; This is the transpose symbol.

[0218] In order to and The deviation minimization of the comprehensive similarity and the deviation between each algorithm result needs to optimize the linear combination coefficient of , and then obtain the optimal , that is: (40)

[0219] In the formula, is the deviation minimization of the comprehensive similarity and the deviation between each algorithm result; is the Euclidean distance (L2 norm); is the optimal weight combination to be solved.

[0220] Using the matrix differential property, the first-order optimality condition is derived:

[0221] (41)

[0222] Solving the linear equation set obtains the optimal linear combination coefficient , and then normalizing it by , and finally obtaining the comprehensive similarity , the formula is:

[0223] (42)

[0224] In the formula, is the comprehensive similarity; is the normalized weight.

[0225] If a node and its neighbor nodes are located in different communities, the node is called a boundary node, so the boundary recheck of the nodes located at the boundary of the community is considered, and the similarity between the node and the subgraph can be measured by the number of internal links at the intersection of the neighborhood of the node and the subgraph. In this embodiment, the definition is used to perform the recheck operation on the boundary node, and the formula is as follows:

[0226] (43)

[0227] Wherein, is the recheck operation on the boundary node, is the boundary node, is the community in which the neighbor node connected to the node is located; is the intersection of the neighbor nodes of and the target community; is the connection weight.

[0228] ​The pruning operation adopts the way of precise deletion, pre-weights the matrix, and performs operation according to the logic of the MCL algorithm, that is, executes the MCL algorithm in the pre-order definition 4. When the iteration process is stopped, the operated matrix is mapped to the community structure. The mapped community structure is subjected to the optimization operation of boundary re-inspection and small community merging. The pseudo code of the project network clustering algorithm is as follows:

[0229]

[0230] In one specific embodiment of the present application, as described in the above step S500, Similar schemes are compared together, errors and workload when a large number of schemes are compared together are reduced, for small sample space, gray target theory is used for decision making, for heterogeneous attributes, are uniformly converted into effect vectors for easy comparison. For clustering decision network, a large number of scheme structures are formed into scheme clusters with similarity, and the gray target decision theory can realize data mining and information development as much as possible.

[0231] Suppose that the scheme cluster has evaluation samples, which constitute the evaluation sample set . and attribute factors constitute the effect sample matrix of , wherein and , are sample elements .

[0232] Since the dimensions and attributes of each index are different, in order to better express the discrete degree between the decision information and the ideal expectation, the reward-punishment transformation operator is introduced, and each index is processed by dimensionless. Let be the average of each attribute factor, then

[0233] (44)

[0234] wherein, is the average value of the attribute ; is the number of schemes; is the scheme index; is the attribute index; is the value of the scheme in the attribute .

[0235] Let be the effect measure of the index , for benefit type index, there is ​

[0236] (45)

[0237] Similarly, for cost-related indicators, we have:

[0238] (46)

[0239] This leads to the normalized decision matrix. .

[0240] For each scheme family, let Then the bullseye, i.e., the optimal effect vector, is ,set up Then the negative bullseye, i.e., the worst-case effect vector, is The distance between the positive and negative targets is... ,plan The positive target center distance is The negative target distance is The evaluation vector of any solution always lies between the positive and negative bullseyes. The optimal solution can be obtained by projecting the distance from the positive bullseye onto the line connecting the positive and negative bullseyes. That is, the larger the projection, the better the corresponding strategy. The projection is the comprehensive bullseye distance. Figure 8 As shown, according to the Law of Cosines, the formula for calculating the projection is:

[0241] (47)

[0242] in, The comprehensive target center distance (projected value) represents the degree of closeness between the proposed scheme and the true target center; This represents the distance between the positive and negative bullseye (global distance). The distance from the target to the bullseye; The distance from the target to the negative bullseye.

[0243] The overall target center distance of the scheme within the cluster and the global value of the scheme are obtained. After scoring (see step S200), the balance between the local and overall aspects of the comprehensive solution is analyzed, and weight values ​​are assigned. This is used to represent the preference of decision-makers regarding the overall and partial performance of a particular option:

[0244] (48)

[0245] in, For weight values, The number of subgroups of solutions. The total number of schemes, For subgroup indexing; The total number of subgroups.

[0246] The final score of the scheme can be expressed as:

[0247] (49)

[0248] wherein, is the final score of the scheme, The value of depends on the sub-group in which the scheme is located, and is calculated by formula (48), is the global value of the scheme, that is, , represents the global performance of the scheme in the entire decision system. Therefore, the final selected scheme is the top schemes with the highest value.

[0249] The method of the present application will be explained in detail below in conjunction with specific examples:

[0250] Among various space propulsion technologies, high-end equipment of heavy-lift launch vehicle engines has the advantages of high performance and reliability, good task adaptability, and many other advantages, and has entered space engineering application first and has always occupied a dominant position. For different application requirements, heavy-lift launch vehicle engines have developed hundreds of engineering products of various types with different thrust levels, different propellants, different propellant supply methods, and different power cycle methods. In the demonstration process of liquid rocket engines under the digital twin scene, multiple enterprises are involved in the design, and each enterprise can provide multiple selection schemes, and the optimal scheme needs to be selected from them. The existing multiple experts evaluate 50 schemes, mainly discuss six attribute indexes of thrust-to-weight ratio, combustion efficiency, variable speed, reliability performance, reuse capability, and complexity, and the specific parameters are shown in Table 1, and the attribute data types are {exact number, interval number, sequence number, language variable, hesitant fuzzy number, probability language term}.

[0251] Table 1 Attribute parameters of heavy-lift launch vehicle engine demonstration

[0252]

[0253] For thrust-to-weight ratio, combustion efficiency, variable speed, etc., which can be output by sensors, the target parameters are set to 1, [0.8, 1.0], {0.80, 0.80, 0.80}. For expert evaluation decision information, the evaluation language variable of reliability performance is mainly three kinds of excellent / good / poor. The hesitant fuzzy number of reuse capability attribute is filled according to the optimistic principle to make its length consistent. The attribute information of complexity is a set of probability language terms of expression form. Part of the evaluation information is shown in Table 2.

[0254] Table 2 Evaluation information of heavy-lift launch vehicle engine demonstration scheme

[0255] ​

[0256] The confidence intervals of characteristic data such as thrust-to-weight ratio, combustion efficiency, and rate of change of the heavy-lift launch vehicle engine are calculated according to formulas (7)-(11), and DU space mapping is performed using formula (12). At the same time, DU space mapping is performed on the heterogeneous information for the evaluation of reliability performance, reusability, and complexity. The R values ​​of the six attributes are calculated using formula (23).

[0257] Based on formulas (23)-(28), the non-compensable attribute values ​​between each pair of schemes are calculated, and the 75% threshold point is obtained, such as... Figure 9 As shown, the red cells represent scheme pairs with a similarity greater than the other schemes.

[0258] Based on the non-compensable attribute values, construct a clustered decision network diagram for the proposed solutions, such as... Figure 10 As shown. Then, the 50 schemes are clustered according to the algorithm. The random state matrix of the traditional MCL algorithm adds a self-loop to each node. When the expansion parameter e=2, the expansion parameter r=2, the maximum number of loops maxloop=60, and the number of self-loops added multfactor=1, the algorithm achieves the best results in dividing the communities. This embodiment sets the parameters as described above, and the experimental results are as follows: the network is divided into 9 groups, namely {3, 4, 7, 10, 45, 14, 47, 48, 20, 31}, {1, 2, 40, 43, 18, 26}, {32, 33, 6, 8, 22, 30}, {42, 11, 44, 21, 23, 27}, {34, 36, 38, 39, 49, 19}, {12, 13, 46, 15, 16}, {0, 35, 37, 17, 25}, {41, 29, 28, 5}, {24, 9}. The clustered network is as follows. Figure 11 As shown, nodes of the same color indicate that they belong to the same cluster. For example, schemes 5, 28 and 41 belong to the same cluster.

[0259] According to gray target decision theory, the effect vectors of all options are first calculated. The cluster of options with the largest number of participants is selected. For example, the effect vectors of option cluster 1 are shown in Table 3:

[0260] Table 3. Effect vector information of scheme cluster 1

[0261]

[0262] The positive and negative target centers of the scheme cluster 1 are calculated as: [0.904, 0.774, -0.130, 0.479, 1.00, 0.454], [-1.000, 0.467, -1.000, -0.613, -0.594, -1.000]

[0263] According to the formula, the positive and negative target center distance is calculated as 3.214. The comprehensive target center distance of the effect vector of each scheme of the scheme cluster 1 is respectively: 2.574, 1.642, 1.921, 1.925, 1.103, 1.751, 1.276, 1.431, 1.469, 1.004.

[0264] According to the TOPSIS score of each scheme in the scheme cluster 1, 0.471, 0.561, 0.473, 0.468, 0.510, 0.497, 0.533, 0.567, 0.462, 0.489, the final score of each scheme in the scheme cluster 1 is taken as 0.5. =0.5, the final score of each scheme in the scheme cluster 1 is 1.522, 1.102, 1.197, 1.196, 0.806, 1.124, 0.905, 0.999, 0.965, 0.747. The final score of the scheme in the remaining scheme cluster can be calculated in the same way, and the final optimal scheme is scheme 23 with a score of 1.47984.

[0265] To sum up, the application constructs a heterogeneous data unification method, maps the data of the digital twin information decision system to the D-H space, converts it into a binary connection number, and describes the certain and uncertain information of the data. Secondly, considering the overall attribute characteristics of high-end equipment, a measure method of attribute non-compensable mechanism is proposed. And the attribute weight is solved by combining the MEREC method. Then, the scheme similarity network is constructed in accordance with the concept of "similar scheme unified evaluation". The node similarity network clustering method of game characteristics is defined to cluster the schemes, and the gray target theory is used for intra-cluster decision-making. Finally, the high-end equipment demonstration case under the digital twin scene is analyzed and discussed in detail. In the experiment, the discrimination degree of the decision-making ranking of the method is high, the time is shortened under the cluster decision-making mode, and the method has an advantage in large-scale schemes. Considering the decision-making scene of heterogeneous data, the accuracy and reliability of the high-end equipment demonstration process are improved.

[0266] According to another aspect of the embodiments of the application, an electronic device is also provided, which includes a processor and a memory, the processor being configured to implement the steps of the method when executing a computer program stored in the memory.

[0267] In the above-mentioned embodiments of the application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0268] In several embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. Among them, the above-mentioned device embodiments are only schematic, for example, the division of the units can be a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, units or modules, and can be electrical or other forms.

[0269] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0270] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various program code storage media.

[0271] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A large-scale heterogeneous multi-attribute equipment scheme clustering decision method, characterized in that, The method comprises the following steps: unifying heterogeneous attribute data in the digital twin decision information system into binary connection numbers; determining attribute weights by using the MEREC method, and constructing attribute irreplaceable values based on the attribute weights and the binary connection numbers; generating a scheme similarity network according to the attribute irreplaceable values; performing network clustering operation on the scheme similarity network, and outputting a scheme clustering result; calculating the comprehensive target center distance of the schemes in the cluster based on the scheme clustering result, fusing the global value of the scheme and the comprehensive target center distance to generate a final scheme score; the method for determining attribute weights by using the MEREC method, and constructing attribute irreplaceable values based on the attribute weights and the binary connection numbers comprises: extracting the certainty component and the uncertainty component in the binary connection number; calculating distance measure, similarity measure, correlation measure and entropy measure based on the certainty component; generating attribute irreplaceable values representing the attribute consistency difference between schemes by weighting and fusing the distance measure, similarity measure, correlation measure and entropy measure through attribute weights; the method for determining attribute weights by using the MEREC method comprises: for benefit-type attributes, normalizing attribute values by using maximum values of each attribute; for cost-type attributes, normalizing attribute values by taking the reciprocal of the minimum value of each attribute; taking the natural logarithm of the normalized attribute values of the benefit-type attributes or the cost-type attributes and summing them up to obtain the total performance value of each scheme; recomputing the total performance value of each scheme after removing each attribute in turn; for each attribute, accumulating the absolute deviation of the original total performance value and the total performance value after removing the attribute; taking the absolute deviation of each attribute and the proportion of the absolute deviation sum of all attributes as the attribute weight value of the attribute.

2. The method of claim 1, wherein, The heterogeneous attribute data comprises sensor measurement information, and the method for converting the sensor measurement information into binary connection numbers comprises: calculating the standardized distance of each scheme in the target database under the same attribute; calculating the positive evidence support degree and the negative evidence support degree of the candidate scheme according to the standardized distance, wherein the positive evidence support degree represents the degree to which the candidate scheme is better than other schemes, and the negative evidence support degree represents the degree to which the candidate scheme is worse than other schemes; constructing a confidence interval based on the positive evidence support degree and the negative evidence support degree, wherein the confidence interval contains the trust degree and the non-negative degree of the attribute of the scheme; mapping the confidence interval to the certainty-uncertainty space, extracting the certainty component and the uncertainty component, and forming the binary connection number.

3. The method of claim 1, wherein, The heterogeneous attribute data comprises expert evaluation information; and the method for converting the expert evaluation information into binary connection numbers comprises the following operations according to the information type: for language variables, quantifying the qualitative language by using a language scale function to generate a quantitative value, calculating the hesitancy degree, and combining the quantitative value and the hesitancy degree to generate a binary connection number; for hesitant fuzzy numbers, calculating the geometric mean hesitancy degree, combining the membership degree score and the geometric mean hesitancy degree to generate a binary connection number; for a probability language term set, calculating the quantitative value of the language term by using a score function, combining the uncertainty represented by a deviation function to generate a binary connection number.

4. The method of claim 1, wherein, The network clustering operation comprises: Adopting game combination optimization strategy, a comprehensive similarity matrix is generated by fusing multiple node similarity calculation methods; Markov clustering operation is performed on the comprehensive similarity matrix, and network expansion, weight inflation and probability pruning are sequentially performed; Boundary nodes in the clustering result are re-inspected for neighborhood link relationship, and the clusters to which the nodes belong are adjusted; Small-sized clusters are merged, and the final scheme clustering result is output.

5. The method of claim 1, wherein, The method for calculating the comprehensive target distance of the schemes in the cluster comprises: Performing reward-punishment transformation on attribute values of the schemes in the cluster, wherein, for benefit-type attributes, an upper limit effect measure is adopted, and the maximum value of the attribute is taken as a benchmark for standardization; for cost-type attributes, a lower limit effect measure is adopted, and the minimum value of the attribute is taken as a benchmark for standardization; Based on the standardized attribute values, a positive target heart and a negative target heart of the schemes in the cluster are determined, wherein the positive target heart is composed of optimal values of the attributes, and the negative target heart is composed of worst values of the attributes; The positive target distance of each scheme from the positive target heart and the negative target distance of each scheme from the negative target heart are calculated; Taking the line connecting the positive and negative target hearts as a benchmark axis, a projection value of the positive target distance of each scheme on the benchmark axis is calculated as the comprehensive target distance.

6. The method of claim 1, wherein, The global value of the scheme is obtained by the following method: Based on the attribute values of all alternative schemes, a positive ideal solution is determined, wherein each attribute takes an optimal value, and for benefit-type attributes, the maximum value is taken, and for cost-type attributes, the minimum value is taken; Based on the attribute values of all alternative schemes, a negative ideal solution is determined, wherein each attribute takes a worst value, and for benefit-type attributes, the minimum value is taken, and for cost-type attributes, the maximum value is taken; For each scheme, the distance between the scheme and the positive ideal solution is calculated, and the distance is obtained by a weighted sum of the difference values of the attribute values, and the weight is the attribute weight; For each scheme, the distance between the scheme and the negative ideal solution is calculated, and the distance is obtained by a weighted sum of the difference values of the attribute values, and the weight is the attribute weight; The global value of the scheme is obtained by dividing the distance between the scheme and the negative ideal solution by the sum of the distance between the scheme and the positive ideal solution and the distance between the scheme and the negative ideal solution.

7. The method of claim 1, wherein, The final scheme score is generated by weighted fusion of the scheme global value and the comprehensive target distance, and the weight coefficient is dynamically determined by the proportion of the scheme subgroup size.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program, when executed by a processor, performs the steps of the large-scale heterogeneous multi-attribute equipment scheme clustering decision method of any one of claims 1-7.

Images