Belief rule base construction method for mesh combat effectiveness evaluation index system

Abstract

The invention provides a belief rule base construction method for a net-shaped combat effectiveness evaluation index system, and the method comprises the steps: setting an evaluation grade, generating an initial rule and a belief rule, determining the relative weight of the belief rule and the transverse influence weight of a lower-layer index, and generating a belief rule base. Quantitative description of a complex influence relationship in a combat effectiveness evaluation index system is realized by constructing a confidence rule base capable of effectively reflecting a longitudinal influence relationship between upper and lower layer indexes and a transverse influence relationship between indexes of the same layer. On the basis of constructing the confidence rule base by using the method provided by the invention, a confidence rule reasoning method can be used for solving a weapon equipment combat effectiveness evaluation problem. According to the method, the evidence reasoning method and the DEMATEL method are combined to construct the confidence rule base suitable for the net-shaped combat effectiveness evaluation index system, and a feasible scheme is provided for expression of knowledge information in the weapon equipment combat effectiveness evaluation problem.

Description

Technical Field

[0001] The invention relates to the field of weapon equipment combat effectiveness evaluation, and in particular to a confidence rule base construction method for a network combat effectiveness evaluation index system. Background Art

[0002] Belief Rule Base (BRB) is a knowledge expression method proposed by Yang Jianbo et al. from the University of Manchester, UK, based on fuzzy rules and combined with evidence theory (Yang JB, Liu J, Wang J, et al. Belief rule-base inference methodology using the evidential reasoning Approach-RIMER[J]. IEEE Transactions on Systems, Man & Cybernetics: Part A, 2006, 36(2): 266-285.). This method can effectively process uncertain information such as incomplete information, fuzzy information, and subjective information, and has been widely used in target recognition (Chinese patent "CN106342322B Aircraft carrier formation recognition method based on confidence rule base reasoning", Chinese patent "CN112800082B A method for identifying aerial targets based on confidence rule base reasoning"), decision evaluation (Chinese patent "CN113033070B A method for monitoring and evaluating pipeline leakage at LNG receiving station terminal", Chinese patent "CN115996082B A traceable airborne bus network health status assessment method and system") and other fields.

[0003] In the field of decision-making evaluation, the existing confidence rule base construction method only focuses on the vertical influence relationship between upper and lower indicators, which can be applied to the traditional tree indicator system. Due to the complexity of the problem of weapon equipment combat effectiveness evaluation, in addition to the vertical influence relationship between upper and lower indicators, there is also a horizontal influence relationship between the same-level indicators. It is necessary to adopt a confidence rule base construction method suitable for the network indicator system, and the existing method is no longer applicable. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides a confidence rule base construction method for a network combat effectiveness evaluation index system, which comprehensively considers the vertical influence relationship between upper and lower layer indicators and the horizontal influence relationship between the same layer indicators.

[0005] The technical solution adopted by the present invention to solve the technical problem comprises the following steps:

[0006] Step 1: Set the evaluation level and define the input and output variables;

[0007] Step 2: Initial rule generation;

[0008] Step 2.1: Convert expert rating information into confidence structure;

[0009] Step 2.2: Generate initial rules;

[0010] Step 3: Generate confidence rules;

[0011] Step 3.1: Initial rule grouping;

[0012] The initial rules with the same premise attributes are grouped into one group, and the P initial rules are grouped into Q groups. The qth group of initial rules is expressed as Where q=1,2,…,Q,r q,t represents the tth initial rule of the qth group, L q represents the number of initial rules in the qth group, t=1,2,…,L q , with the relationship:

[0013]

[0014] Step 3.2: Calculate the initial rule similarity;

[0015] Calculate the similarity of the initial rules. The similarity of the tth initial rule in the qth group is:

[0016]

[0017] Among them, γ q,t represents the tth initial rule r of the qth group q,t The similarity of represents the i-th lower-level index U of the t-th initial rule in the q-th group i confidence level;

[0018] Step 3.3: Initial rule aggregation;

[0019] The initial rules in the same group are aggregated using the evidence reasoning method to generate confidence rules. According to equations (9) and (10), L in the qth group is q Aggregate the initial rules to generate the qth confidence rule;

[0020] Similarly, the other groups of initial rules obtained in step 3.1 are calculated through steps 3.2 to 3.3, and the Q groups of initial rules generate a total of Q confidence rules;

[0021] Step 4: Determine the relative weights of the confidence rules;

[0022] Step 4.1: Calculate the importance of the confidence rule;

[0023] The qth confidence rule is generated from the qth group of initial rules. The importance of this confidence rule is:

[0024]

[0025] Among them, D q Indicates the importance of the qth confidence rule; L q represents the number of initial rules in the qth group; γ q,t represents the similarity of the t-th initial rule in the q-th group;

[0026] Step 4.2: Calculate the relative weights of the confidence rules;

[0027] Calculate the relative weight of the confidence rules, the qth confidence rule R q The relative weights are:

[0028]

[0029] Among them, θ q represents the relative weight of the qth confidence rule; D q Indicates the importance of the qth confidence rule; max{D i |i=1,2,…,Q} represents the maximum value of the confidence rule importance;

[0030] Step 5: Determine the horizontal impact weight of the lower-level indicators;

[0031] The DEMATEL method is used to determine the horizontal impact weights of the lower-level indicators;

[0032] Step 6: Generate a confidence rule base;

[0033] The confidence rules, relative weights and horizontal impact weights are combined to generate a confidence rule base for the network combat effectiveness evaluation index system, where the qth confidence rule is expressed as:

[0034]

[0035] Among them, R q represents the qth confidence rule in the confidence rule base; U i represents the i-th lower-level indicator of the confidence rule base; d q,i represents the i-th lower level indicator U in the q-th confidence rule i Corresponding reference value; N u Indicates the number of lower-level indicators; It represents the jth evaluation level of the combat effectiveness index A; represents the jth evaluation level of the combat effectiveness index A in the qth confidence rule Confidence level; M AIndicates the number of evaluation levels of the upper-level indicator A; θ q represents the relative weight of the qth confidence rule in the confidence rule base; δ i Represents the i-th premise attribute U i The horizontal impact weight of

[0036] By constructing a confidence rule base that can effectively reflect the vertical influence relationship between upper and lower level indicators and the horizontal influence relationship between the same level indicators, a quantitative description of the complex influence relationship in the combat effectiveness evaluation indicator system is achieved. On the basis of using the confidence rule base constructed by the method proposed by the present invention, the confidence rule reasoning method can be used to solve the problem of weapon equipment combat effectiveness evaluation.

[0037] In step 1, in the operational effectiveness evaluation index system, the operational effectiveness index A and the lower-level index set There is a vertical influence relationship between U i Represents the i-th index in the lower layer index, N U Indicates the number of lower-level indicators, i = 1, 2, ..., N U ; Lower level indicators There is a horizontal influence relationship between them, which can be expressed by the influence relationship matrix shown in formula (1):

[0038]

[0039] Among them, C represents the influence relationship matrix, c m,n The element in the mth row and nth column of the influence relationship matrix C represents the mth lower-level indicator U m For the nth lower level index U n The degree of influence, m = 1, 2, ..., N U ,n=1,2,…,N U ;

[0040] The evaluation level of combat effectiveness index A is set to The reference values ​​for each level are in, represents the jth evaluation level of the combat effectiveness indicator A; Indicates the evaluation level Corresponding reference value; M A Represents the number of evaluation levels of combat effectiveness index A, j = 1, 2, ..., M A ;

[0041] Set the i-th lower layer index U i The evaluation level is The reference values ​​for each level are in, represents the i-th lower layer index U iThe kth evaluation level of ; represents the i-th lower layer index U i kth evaluation level Corresponding reference value; M i represents the i-th lower layer index U i The number of evaluation levels, k = 1, 2, ..., M i ;

[0042] Define the lower level indicator set {U i |i=1,2,…,N U} is the input variable of the confidence rule base, defining the confidence structure information of the combat effectiveness index A is the output variable of the confidence rule base;

[0043] In step 2.1, the steps of converting the expert scoring information into a confidence structure are:

[0044] Through the expert scoring method, a total of P expert scoring information is obtained, and the p-th expert scoring information is expressed as in, represents the score value set of the lower-level indicators in the p-th expert scoring information, x p,i Indicates the p-th expert scoring information for the i-th lower-level indicator U i The scoring value of x p,A represents the score value of the combat effectiveness index A in the p-th expert scoring information;

[0045] According to formula (2), the scoring value of the lower-level indicator is converted into a confidence structure:

[0046]

[0047] in, represents the i-th lower layer index U i The kth evaluation level of The corresponding reference value; Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value of its k-th evaluation level confidence level;

[0048] The i-th lower-level indicator U in the p-th expert scoring information i The score value x p,i After being converted into a confidence structure, it is expressed as:

[0049]

[0050] Among them, X p,i Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value xp,i The confidence structure information obtained by transformation;

[0051] Similarly, the score value x of the combat effectiveness index A in the pth expert scoring information is p,A Transformed into a confidence structure, expressed as:

[0052]

[0053] Among them, X p,A represents the score value x of the combat effectiveness index A in the pth expert scoring information p,A The confidence structure information obtained by transformation; It represents the score value of the combat effectiveness index A in the p-th expert scoring information and its j-th evaluation level confidence level.

[0054] In the step 2.2, initial rules are generated;

[0055] Generate the initial rules according to formula (5):

[0056]

[0057] in, Represents the i-th lower-level indicator U in the p-th expert scoring information i The confidence structure information X p,i The maximum value of medium confidence; Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value of its k-th evaluation level confidence level;

[0058] Based on the pth expert scoring information x p The generated initial rules are expressed as:

[0059]

[0060] Among them, r p represents the pth initial rule; U i represents the i-th lower-level indicator in the vertical influence relationship; Represents the i-th lower-level indicator U in the p-th expert scoring information i The confidence structure information X p,i The reference value corresponding to the evaluation level with the largest confidence is defined as the i-th premise attribute of the p-th initial rule; Indicates the reference value of the assessment level confidence level; represents the jth evaluation level of the combat effectiveness indicator A; It represents the score value of the combat effectiveness index A in the p-th expert scoring information and its j-th evaluation level confidence level;

[0061] The collected P expert scoring information is calculated to obtain P initial rules.

[0062] In step 3.3, the qth confidence rule is generated:

[0063]

[0064]

[0065] in, represents the jth evaluation level of the combat effectiveness indicator A in the qth confidence rule Confidence of γ q,t represents the similarity of the t-th initial rule in the q-th group; represents the jth evaluation level of the combat effectiveness indicator A in the tth initial rule of the qth group Confidence level; d q represents the computational intermediate amount required to aggregate the qth group of initial rules using the evidential reasoning method;

[0066] The generated qth confidence rule is expressed as:

[0067]

[0068] Among them, R q represents the qth confidence rule in the confidence rule base; U i represents the i-th lower-level indicator of the confidence rule base; d q,i represents the i-th lower level indicator U in the q-th confidence rule i The corresponding reference value; Represents the j-th evaluation level of the combat effectiveness indicator A.

[0069] In step 5, the DEMATEL method is used to determine the horizontal impact weight of the lower-level indicators;

[0070] First, according to equations (14) and (15), the influence relationship matrix C is transformed into the normalized influence matrix Z:

[0071]

[0072]

[0073] Among them, c m,n Represents the element in the mth row and nth column of the influence relationship matrix C, z m,nRepresents the element in the mth row and nth column of the normalized influence matrix Z, where m,n=1,2,…,N U ;

[0074] Then, the comprehensive influence matrix T is calculated according to formula (16), and the influence vector is calculated according to formula (17) and formula (18) respectively. and the influence vector

[0075] T=Z(IZ) -1 (16)

[0076]

[0077]

[0078] Where T represents the comprehensive influence matrix; I represents the identity matrix with the same size as the normalized influence matrix Z; H m represents the mth element in the influence vector H; F n represents the nth element in the influence vector F; t m,n represents the element in the mth row and nth column of the comprehensive influence matrix T;

[0079] Finally, the i-th lower layer index U is calculated i The horizontal impact weight is:

[0080]

[0081] Among them, δ i represents the i-th lower layer index U i The horizontal impact weight of H i represents the i-th element in the influence vector H; F i Represents the i-th element in the influence vector F;

[0082] Calculate the horizontal impact weight of each lower-level indicator and form a horizontal impact weight set

[0083] The beneficial effect of the present invention is that the existing confidence rule base construction method in the field of decision evaluation only focuses on the vertical influence relationship between upper and lower layer indicators, is only applicable to tree indicator systems, and cannot be applicable to network combat effectiveness evaluation indicator systems. The present invention comprehensively considers the vertical influence relationship between upper and lower layer indicators and the horizontal influence relationship between the same layer indicators, combines the evidence reasoning method and the DEMATEL method to realize the construction of a confidence rule base applicable to the network combat effectiveness evaluation indicator system, and provides a feasible solution for the expression of knowledge information in the problem of weapon equipment combat effectiveness evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 It is a schematic diagram of the network combat effectiveness evaluation index system in a specific implementation manner of the present invention. DETAILED DESCRIPTION

[0085] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0086] Figure 1 The figure shows a typical network combat effectiveness evaluation index system, in which the vertical influence relationship is represented by a solid arrow, specifically the influence relationship between the combat effectiveness index A and the lower-level indicators of reconnaissance and perception capability U1, command and decision-making capability U2, coordinated combat capability U3 and damage assessment capability U4; the horizontal influence relationship is represented by a dotted arrow, specifically the influence of reconnaissance and perception capability U1 on command and decision-making capability U2, reconnaissance and perception capability U1 on coordinated combat capability U3, reconnaissance and perception capability U1 on damage assessment capability U4, and command and decision-making capability U2 on coordinated combat capability U3, constructing the influence relationship matrix C:

[0087]

[0088] For Figure 1 The network combat effectiveness evaluation index system shown in the figure builds a trust rule base. The steps are as follows:

[0089] Step 1: Set the evaluation level and define the input and output variables

[0090] The evaluation level of each indicator in the evaluation index system is set to {D1, D2, D3, D4, D5}, which represent low, lower, medium, higher and high respectively, and the utility value of each level is {d1=0, d2=0.25, d3=0.5, d4=0.75, d5=1}.

[0091] in accordance with Figure 1 The vertical influence relationship of the evaluation index system shown in the figure sets the lower-level indicators of reconnaissance and perception capability U1, command and decision-making capability U2, coordinated combat capability U3 and damage assessment capability U4 as the input of the confidence rule base, and the confidence structure information of the combat effectiveness index A as the output of the confidence rule base.

[0092] Step 2: Initial rule generation

[0093] Step 2.1: Convert expert rating information into confidence structure;

[0094] 9600 expert scoring information was obtained through the expert scoring method, as shown in Table 1.

[0095] Table 1 Expert scoring information

[0096]

[0097] According to formula (2), the expert scoring information in Table 1 is converted into a confidence structure, as shown in Table 2.

[0098] Table 2 Expert scoring information represented by confidence structure

[0099]

[0100]

[0101] Step 2.2: Generate initial rules;

[0102] According to formula (5), the expert scoring information represented by the confidence structure in Table 2 is converted into initial rules, as shown in Table 3:

[0103] Table 3 Initial rule list

[0104]

[0105] Step 3: Generate confidence rules;

[0106] Step 3.1: Initial rule grouping;

[0107] In the initial rules in Table 3, the 1st, 2nd and 4th initial rules have the same premise attributes and are grouped into one group. Similarly, the other initial rules in Table 3 are grouped into 3125 groups in total.

[0108] Step 3.2: Calculate the initial rule similarity;

[0109] The following is an example of the initial rules 1, 2, and 4.

[0110] According to formula (8), the similarities of the first, second and fourth initial rules are 0.6006, 0.5109 and 0.7004 respectively.

[0111] Step 3.3: Initial rule aggregation;

[0112] The first, second and fourth initial rules are fused using equations (9) and (10) to obtain a confidence rule:

[0113]

[0114] Similarly, the similarities of the initial rules of other groups are calculated and the initial rules are aggregated, and a total of 3125 confidence rules are obtained.

[0115] Step 4: Determine the relative weights of the confidence rules

[0116] Step 4.1: Calculate the importance of the confidence rule;

[0117] The importance of each confidence rule is calculated according to formula (12). The importance of the confidence rule shown in formula (22) is 0.6006+0.5109+0.7004=1.8119. Among all the confidence rules, the maximum importance is 2.0223.

[0118] Step 4.2: Calculate the relative weights of the confidence rules;

[0119] According to formula (13), the relative weight of the confidence rule shown in formula (20) is calculated to be 1.8119 / 2.0223=0.8960.

[0120] Similarly, the relative weight of each confidence rule is calculated.

[0121] Step 5: Determine the horizontal impact weight of the lower-level indicators

[0122] According to formula (15), the normalized influence matrix Z is calculated as:

[0123]

[0124] According to formula (16), the comprehensive influence matrix T is calculated as follows:

[0125]

[0126] According to formula (17), the influence vector is calculated as H = [0, 0.3333, 0.7778, 0, 3333].

[0127] According to formula (18), the influence vector is calculated as F = [1.1111, 0.3333, 0, 0].

[0128] Finally, according to formula (19), the lateral influence weights of the lower-level indicators U1, U2, U3 and U4 are calculated to be 0.3846, 0.2308, 0.2692 and 0.1154 respectively.

[0129] Step 6: Generate a confidence rule base

[0130] Combine the above confidence rules, relative weights and horizontal impact weights to generate Figure 1 The confidence rule base of the network combat effectiveness evaluation index system is shown in Table 4:

[0131] Table 4 Confidence rule base

[0132]

[0133] As can be seen from Table 4, the confidence rule base contains a total of 3125 confidence rules.

Claims

1. A method for constructing a trust rule base for a network combat effectiveness evaluation index system, characterized in that The steps include: Step 1: Set the evaluation level and define the input and output variables; Step 2: Initial rule generation; Step 2.1: Convert expert rating information into confidence structure; Step 2.2: Generate initial rules; Step 3: Generate confidence rules; Step 3.1: Initial rule grouping; The initial rules with the same premise attributes are grouped into one group, and the P initial rules are grouped into Q groups. The qth group of initial rules is expressed as Where q=1,2,…,Q,r q,t represents the tth initial rule of the qth group, L q represents the number of initial rules in the qth group, t=1,2,…,L q , with the relationship: Step 3.2: Calculate the initial rule similarity; Calculate the similarity of the initial rules. The similarity of the tth initial rule in the qth group is: Among them, γ q,t represents the tth initial rule r of the qth group q,t The similarity of represents the i-th lower-level index U of the t-th initial rule in the q-th group i confidence level; Step 3.3: Initial rule aggregation; The initial rules in the same group are aggregated using the evidence reasoning method to generate confidence rules. According to equations (9) and (10), L in the qth group is q Aggregate the initial rules to generate the qth confidence rule; Similarly, the other groups of initial rules obtained in step 3.1 are calculated through steps 3.2 to 3.3, and the Q groups of initial rules generate a total of Q confidence rules; Step 4: Determine the relative weights of the confidence rules; Step 4.1: Calculate the importance of the confidence rule; The qth confidence rule is generated from the qth group of initial rules. The importance of this confidence rule is: Among them, D q Indicates the importance of the qth confidence rule; L q represents the number of initial rules in the qth group; γ q,t represents the similarity of the tth initial rule in the qth group; Step 4.2: Calculate the relative weights of the confidence rules; Calculate the relative weight of the confidence rules, the qth confidence rule R q The relative weights are: Among them, θ q represents the relative weight of the qth confidence rule; D q Indicates the importance of the qth confidence rule; max{D i |i=1,2,…,Q} represents the maximum value of the confidence rule importance; Step 5: Determine the horizontal impact weight of the lower-level indicators; The DEMATEL method is used to determine the horizontal impact weights of the lower-level indicators; Step 6: Generate a confidence rule base; The confidence rules, relative weights and horizontal impact weights are combined to generate a confidence rule base for the network combat effectiveness evaluation index system, where the qth confidence rule is expressed as: Among them, R q represents the qth confidence rule in the confidence rule base; U i represents the i-th lower-level indicator of the confidence rule base; d q,i represents the i-th lower level indicator U in the q-th confidence rule i Corresponding reference value; N u Indicates the number of lower-level indicators; It represents the jth evaluation level of the combat effectiveness index A; represents the jth evaluation level of the combat effectiveness index A in the qth confidence rule Confidence level; M A Indicates the number of evaluation levels of the upper-level indicator A; θ q represents the relative weight of the qth confidence rule in the confidence rule base; δ i Represents the i-th premise attribute U i The horizontal impact weight of By constructing a confidence rule base that can effectively reflect the vertical influence relationship between upper and lower level indicators and the horizontal influence relationship between the same level indicators, a quantitative description of the complex influence relationship in the combat effectiveness evaluation index system is achieved. On the basis of constructing the confidence rule base using the method proposed by the present invention, the confidence rule reasoning method can be used to solve the problem of weapon equipment combat effectiveness evaluation.

2. The method for constructing a trust rule base for a networked combat effectiveness evaluation index system according to claim 1, characterized in that: In step 1, in the operational effectiveness evaluation index system, the operational effectiveness index A and the lower-level index set There is a vertical influence relationship between U i Represents the i-th index in the lower layer index, N U Indicates the number of lower-level indicators, i = 1, 2, ..., N U ; Lower level indicators There is a horizontal influence relationship between them, which can be expressed by the influence relationship matrix shown in formula (1): Among them, C represents the influence relationship matrix, c m,n The element in the mth row and nth column of the influence relationship matrix C represents the mth lower-level indicator U m For the nth lower level index U n The degree of influence, m = 1, 2, ..., N U ,n=1,2,…,N U ; The evaluation level of combat effectiveness index A is set to The reference values ​​for each level are in, represents the jth evaluation level of the combat effectiveness indicator A; Indicates the evaluation level Corresponding reference value; M A Represents the number of evaluation levels of combat effectiveness index A, j = 1, 2, ..., M A ; Set the i-th lower layer index U i The evaluation level is The reference values ​​for each level are in, represents the i-th lower layer index U i The kth evaluation level of ; represents the i-th lower layer index U i kth evaluation level Corresponding reference value; M i represents the i-th lower layer index U i The number of evaluation levels, k = 1, 2, ..., M i ; Define the lower level indicator set {U i |i=1,2,…,N U } is the input variable of the confidence rule base, defining the confidence structure information of the combat effectiveness index A is the output variable of the confidence rule base.

3. The method for constructing a trust rule base for a networked combat effectiveness evaluation index system according to claim 1, characterized in that: In step 2.1, the steps of converting the expert scoring information into a confidence structure are: Through the expert scoring method, a total of P expert scoring information is obtained, and the pth expert scoring information is expressed as x p ={[x p,1 ,x p,2 ,…,x p,i ,…,x p,NU ],x p,A },in, represents the score value set of the lower-level indicators in the p-th expert scoring information, x p,i Indicates the p-th expert scoring information for the i-th lower-level indicator U i The scoring value of x p,A represents the score value of the combat effectiveness index A in the p-th expert scoring information; According to formula (2), the scoring value of the lower-level indicator is converted into a confidence structure: in, represents the i-th lower layer index U i The kth evaluation level of The corresponding reference value; Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value of its k-th evaluation level confidence level; The i-th lower-level indicator U in the p-th expert scoring information i The score value x p,i After being converted into a confidence structure, it is expressed as: Among them, X p,i Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value x p,i The confidence structure information obtained by transformation; Similarly, the score value x of the combat effectiveness index A in the pth expert scoring information is p,A Transformed into a confidence structure, expressed as: Among them, X p,A represents the score value x of the combat effectiveness index A in the pth expert scoring information p,A The confidence structure information obtained by transformation; It represents the score value of the combat effectiveness index A in the p-th expert scoring information and its j-th evaluation level confidence level.

4. The method for constructing a trust rule base for a networked combat effectiveness evaluation index system according to claim 1, characterized in that: In the step 2.2, initial rules are generated; Generate the initial rules according to formula (5): in, Represents the i-th lower-level indicator U in the p-th expert scoring information i The confidence structure information X p,i The maximum value of medium confidence; Represents the i-th lower-level indicator U in the p-th expert scoring information i The score value of its k-th evaluation level confidence level; Based on the pth expert scoring information x p The generated initial rules are expressed as: Among them, r p represents the pth initial rule; U i represents the i-th lower-level indicator in the vertical influence relationship; Represents the i-th lower-level indicator U in the p-th expert scoring information i The confidence structure information X p,i The reference value corresponding to the evaluation level with the largest confidence is defined as the i-th premise attribute of the p-th initial rule; Indicates the reference value of the assessment level confidence level; represents the jth evaluation level of the combat effectiveness indicator A; It represents the score value of the combat effectiveness index A in the p-th expert scoring information and its j-th evaluation level confidence level; The collected P expert scoring information is calculated to obtain P initial rules.

5. The method for constructing a trust rule base for a networked combat effectiveness evaluation index system according to claim 1, characterized in that: In step 3.3, the qth confidence rule is generated: in, represents the jth evaluation level of the combat effectiveness indicator A in the qth confidence rule Confidence of q,t represents the similarity of the tth initial rule in the qth group; represents the jth evaluation level of the combat effectiveness indicator A in the tth initial rule of the qth group Confidence level; d q represents the computational intermediate amount required to aggregate the qth group of initial rules using the evidential reasoning method; The generated qth confidence rule is expressed as: Among them, R q represents the qth confidence rule in the confidence rule base; U i represents the i-th lower-level indicator of the confidence rule base; d q,i represents the i-th lower level indicator U in the q-th confidence rule i The corresponding reference value; Represents the j-th evaluation level of the combat effectiveness indicator A.

6. The method for constructing a trust rule base for a networked combat effectiveness evaluation index system according to claim 1, characterized in that: In step 5, the DEMATEL method is used to determine the horizontal impact weight of the lower-level indicators; First, according to equations (14) and (15), the influence relationship matrix C is transformed into the normalized influence matrix Z: Among them, c m,n Represents the element in the mth row and nth column of the influence relationship matrix C, z m,n Represents the element in the mth row and nth column of the normalized influence matrix Z, where m,n=1,2,…,N U ; Then, the comprehensive influence matrix T is calculated according to formula (16), and the influence vector is calculated according to formula (17) and formula (18) respectively. and the influence vector T=Z(I-Z) -1 (16) Where T represents the comprehensive influence matrix; I represents the identity matrix with the same size as the normalized influence matrix Z; H m represents the mth element in the influence vector H; F n represents the nth element in the influence vector F; t m,n represents the element in the mth row and nth column of the comprehensive influence matrix T; Finally, the i-th lower layer index U is calculated i The horizontal impact weight is: Among them, δ i represents the i-th lower layer index U i The horizontal impact weight of H i represents the i-th element in the influence vector H; F i Represents the i-th element in the influence vector F; Calculate the horizontal impact weight of each lower-level indicator and form a horizontal impact weight set

Images