A power optical transmission network diagnosis method based on cloud model evidence cloud drop mechanism

Abstract

The present application relates to a kind of power optical transmission network diagnosis method based on cloud model evidence cloud drop mechanism, comprising the following steps: determining each index for diagnosing power optical transmission network, index includes qualitative index and quantitative index;Determine index weight;Determine the reliability grade cloud model of qualitative index, obtain the comprehensive cloud model of qualitative index according to reliability grade cloud model, obtain the grade support degree of qualitative index according to comprehensive cloud model, and calculate the grade membership of qualitative index, with maximum membership as the diagnosis result of qualitative index;According to the diagnosis result of quantitative index and qualitative index, the power optical transmission network is diagnosed.The present application realizes the conversion between qualitative and quantitative by cloud model, then utilizes cloud generator to generate representative cloud drop, and carries out comprehensive analysis to the mutual influence between cloud drop by combining evidence theory, obtains the grade support degree of qualitative index to each grade.The general comprehensive diagnosis conclusion of the reliability of this power optical transmission network is obtained by comprehensively analyzing two types of indexes.

Description

Technical Field

[0001] This invention relates to a diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism, belonging to the field of power optical transmission network maintenance technology. Background Technology

[0002] Power optical transmission network diagnosis refers to the process of qualitatively and quantitatively analyzing the technology and economics of power optical transmission communication networks under an objective and comprehensive indicator system. This involves monitoring the phased changes of medium- and long-term node indicators, combining development needs and actual operational conditions, determining dynamic judgment conditions, and employing scientific and comprehensive judgment methods to form diagnostic conclusions and provide development recommendations. Current power optical transmission network diagnosis methods typically assign weights to each quantitative and qualitative indicator after determination, and then simply assign amplitude values ​​to the qualitative indicators based on expert opinions. This leads to significant subjective influence on the diagnostic results. Summary of the Invention

[0003] To overcome the aforementioned problems, this invention provides a diagnostic method for power optical transmission networks based on a cloud model-based evidence cloud droplet mechanism. This method utilizes a cloud model to convert between qualitative and quantitative methods. Then, a cloud generator is used to produce representative cloud droplets, and the mutual influence between cloud droplets is comprehensively analyzed using evidence theory to obtain the support degree of qualitative indicators for each reliability level. Finally, the membership degrees of the two types of indicators to each reliability level are combined to obtain a general comprehensive diagnostic conclusion on the reliability of the power optical transmission network.

[0004] The technical solution of the present invention is as follows:

[0005] First aspect

[0006] A diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism includes the following steps:

[0007] Determine the various indicators used for diagnosing power optical transmission networks, including qualitative and quantitative indicators;

[0008] Determine the weights of the indicators;

[0009] A reliability level cloud model for qualitative indicators is determined. A comprehensive cloud model for qualitative indicators is obtained based on the reliability level cloud model. The level support degree of qualitative indicators is obtained based on the comprehensive cloud model. The level membership degree of qualitative indicators is calculated, and the maximum membership degree is taken as the diagnostic result of the qualitative indicator.

[0010] The power optical transmission network is diagnosed based on the diagnostic results of quantitative and qualitative indicators.

[0011] Furthermore, the qualitative indicators include the completeness of spare parts and the scalability of the network management system, while the quantitative indicators include the aging rate of optical cables, the average utilization rate of optical cores, and the average configuration of node equipment.

[0012] Furthermore, the determination of the indicator weights specifically involves:

[0013] The indicators were compared pairwise using a scaling method, specifically by experts who assessed their importance. Indicator a was assigned a rating of 1, 3, 5, 7, and 9 relative to indicator b, respectively, indicating it was equally important, slightly important, significantly important, strongly important, and extremely important. The results were then presented in the following matrix:

[0014]

[0015] in, and Let A and J represent the upper and lower limits of the relative reliability comparison results of the i-th and j-th indicators, respectively. According to the interval number operation method, we have A = [A...]. - A + ], where A + A - It is a matrix of upper and lower bounds;

[0016] Calculate A using the characteristic root method for interval numbers. + A - The corresponding largest eigenvalues And the normalized feature w with positive components + w - ;

[0017] Calculate the positive component coefficients m and negative component coefficients k of matrix A:

[0018]

[0019] Where n is the number of indicators;

[0020] Calculate matrix A + A - The corresponding consistency test index values ​​CR + CR - ,when When the value is less than 0.1, the consistency test is passed, and the interval weight vector of the indicator is obtained as W = [kw - ,mw + ].

[0021] Furthermore, the cloud model for determining the reliability level of qualitative indicators specifically refers to:

[0022] The qualitative indicators are divided into N reliability levels;

[0023] Let C ij =(Ex ij En ij He ij ) represents the reliability level cloud model corresponding to qualitative index i, where Ci1 C iN and C ip The corresponding reliability level cloud models are semi-falling cloud, semi-rising cloud, and normal cloud, respectively; p = 2, 3, ..., N-1;

[0024] Furthermore, the digital characteristics of the reliability level cloud are calculated using the following method:

[0025] For semi-falling clouds C i1 :

[0026] Ex ij =a i1 ;

[0027] En ij =(b i1 -a i1 ) / 3;

[0028] He ij =ε i1 ;

[0029] For normal cloud C ip :

[0030] Ex ij =(b ij +a ij ) / 2;

[0031] En ij =(b ij -a ij ) / 6;

[0032] He ij =ε ij ;

[0033] Where j = p;

[0034] For half-liter cloud C iN :

[0035] Ex ij =b iN ;

[0036] En ij =(b iN -a iN ) / 3;

[0037] He ij =ε iN ;

[0038] Where, ε i1 ε ij and ε iN It is a constant.

[0039] Furthermore, based on the aforementioned reliability level cloud model, a comprehensive cloud model with qualitative indicators is obtained, specifically as follows:

[0040] Obtain the reliability level cloud model C of the k-th expert on the i-th qualitative indicator. ik ;

[0041] The comprehensive cloud model C for calculating the i-th qualitative indicator. i =(Ex j En j He j ):

[0042]

[0043] Among them, Ex i En i He i The integrated cloud model C is respectively i Expectation, entropy, hyperentropy; λ k Let be the weight of the k-th expert, k = 1, 2, 3, ..., t, where t is the number of experts; Ex ik En ik He ik These represent the expected value, entropy, and hyperentropy of the cloud model for the reliability level of the i-th qualitative indicator by the k-th expert.

[0044] Furthermore, based on the comprehensive cloud model, the grade support of the qualitative indicators is obtained, specifically as follows:

[0045] Determine the distance between the integrated cloud model and the cloud models of each reliability level, specifically by generating the integrated cloud model C using a forward cloud generator. i Given T evidence cloud droplets, the reliability level cloud model C of evidence cloud droplet m is calculated using the following formula. j Center (Ex) j Euclidean geometric distance of (,0):

[0046]

[0047] Calculate the basic probability assignment function for each level of evidence cloud droplet m:

[0048]

[0049] Where, m im (A j (C) is a comprehensive cloud model. i The basic probability assignment function of evidence cloud drop m to level j, i.e. the probability that evidence cloud drop m supports level j;

[0050] Calculate the divergence of evidence cloud droplet m:

[0051]

[0052] Among them, KL(m im ||m il ) represents the divergence of evidence cloud drop m, indicating the degree of closeness between evidence cloud drop m and evidence cloud drop l in terms of support at each level. The smaller the value, the higher the degree of closeness.

[0053] Calculate the credibility of evidence cloud droplet m:

[0054]

[0055] The credibility of the evidence cloud droplets was normalized:

[0056]

[0057] The normalized confidence level ω of T evidence cloud drops is obtained. im =[ω i1 ,ω i2 ,...,ω iT ] T ;

[0058] Based on the normalized confidence level of the evidence cloud droplets, determine the comprehensive probability assignment function of evidence cloud droplet m to level j;

[0059]

[0060] According to the DS evidence theory composition rules, calculate the support of the i-th qualitative indicator for grade j:

[0061]

[0062] in, This refers to the conflict factor in the synthesis formula of evidence theory.

[0063] Furthermore, the calculation of the membership degree of the qualitative indicators is as follows:

[0064]

[0065]

[0066] in, This is the sum of the correlations between all qualitative indicators and the optimal solution set at level j; This is the sum of the correlations between all qualitative indicators and the worst-case solution set of level j.

[0067] Calculate the membership degree of all qualitative indicators relative to each level:

[0068]

[0069] The overall membership degree of this power optical transmission network to each level is obtained:

[0070] K j (U)=K j1 (U)+K j2 (U).

[0071] Second aspect

[0072] A diagnostic system for power optical transmission networks based on cloud model evidence cloud droplet mechanism includes a data input module and a calculation module. The data input module inputs the numerical values ​​of indicators for diagnosing the power optical transmission network and the expert's evaluation results on the importance of the indicators. The calculation module calculates and outputs the comprehensive membership degree of the power optical transmission network to each level according to the method described in the first aspect.

[0073] Third aspect

[0074] A storage medium storing a program that, after being executed by a computer, outputs the comprehensive membership degree of the power optical transmission network to each level according to the method described in the first aspect.

[0075] The present invention has the following beneficial effects:

[0076] This method utilizes a cloud model to convert between qualitative and quantitative methods. Then, a cloud generator produces representative cloud droplets, and evidence theory is used to comprehensively analyze the interactions between these droplets, yielding the support degree of qualitative indicators for each reliability level. Finally, the membership degrees of both types of indicators to each reliability level are combined to obtain a general comprehensive diagnostic conclusion for the reliability of the power optical transmission network. Attached Figure Description

[0077] Figure 1 This is a flowchart of the method of the present invention.

[0078] Figures 2(a)-(b) are cross-sectional views of the present invention. Detailed Implementation

[0079] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0080] First aspect

[0081] Example 1

[0082] A diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism includes the following steps:

[0083] Determine the various indicators used for diagnosing power optical transmission networks, including qualitative and quantitative indicators;

[0084] Determine the weights of the indicators;

[0085] A reliability level cloud model for qualitative indicators is determined. A comprehensive cloud model for qualitative indicators is obtained based on the reliability level cloud model. The level support degree of qualitative indicators is obtained based on the comprehensive cloud model. The level membership degree of qualitative indicators is calculated, and the maximum membership degree is taken as the diagnostic result of the qualitative indicator.

[0086] The power optical transmission network is diagnosed based on the diagnostic results of quantitative and qualitative indicators.

[0087] In one specific embodiment of the present invention, the qualitative indicators include the completeness of spare parts and the scalability of the network management system, and the quantitative indicators include the aging rate of optical cables, the average utilization rate of optical cores, and the average configuration of node equipment.

[0088] Example 2

[0089] A diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism, in addition to the method described in Example 1, specifically involves determining the weights of the indicators as follows:

[0090] The indicators were compared pairwise using the scaling method. Specifically, the importance was assessed by experts. Indicator a was assigned a level of 1, 3, 5, 7, and 9 relative to indicator b, respectively, as shown in Table 1.

[0091] Table 1 Scale Rules

[0092] grade Degree of language description meaning 1 Equal Equal importance of indicator a and indicator b 3 a little Indicator A is slightly more important than indicator B. 5 obvious Indicator A is significantly more important than indicator B. 7 strong Indicator A is significantly more important than Indicator B 9 extreme Indicator A is more extremely important than indicator B

[0093] And form the following matrix:

[0094]

[0095] in, and Let A and J represent the upper and lower limits of the relative reliability comparison results of the i-th and j-th indicators, respectively. According to the interval number operation method, we have A = [A...]. - A + ], where A + A - It is a matrix of upper and lower bounds;

[0096] Calculate A using the interval number eigenvalue method (IEM). + A - The corresponding largest eigenvalues And the normalized feature w with positive components + w - ;

[0097] Calculate the positive component coefficients m and negative component coefficients k of matrix A:

[0098]

[0099] Where n is the number of indicators;

[0100] Calculate matrix A + A - The corresponding consistency test index values ​​CR + CR - ,when When the value is less than 0.1, the consistency test is passed, and the interval weight vector of the indicator is obtained as W = [kw - ,mw + ].

[0101] Example 3

[0102] A diagnostic method for power optical transmission networks based on a cloud model evidence cloud droplet mechanism, in addition to the method described in Example 1, specifically defines the cloud model for determining the reliability level of qualitative indicators as follows:

[0103] The qualitative indicators are divided into N reliability levels;

[0104] Let C ij =(Ex ij En ij He ij ) represents the reliability level cloud model corresponding to qualitative index i, where C i1 C iN and C ip The corresponding reliability level cloud models are semi-falling cloud, semi-rising cloud, and normal cloud, respectively; p = 2, 3, ..., N-1;

[0105] In one embodiment of this disclosure, the digital characteristics of the reliability level cloud are calculated according to the following method:

[0106] For semi-falling clouds C i1 :

[0107] Ex ij =a i1 ;

[0108] En ij =(b i1 -a i1 ) / 3;

[0109] He ij =ε i1 ;

[0110] For normal cloud C ip :

[0111] Ex ij =(b ij+a ij ) / 2;

[0112] En ij =(b ij -a ij ) / 6;

[0113] He ij =ε ij ;

[0114] Where j = p;

[0115] For half-liter cloud C iN :

[0116] Ex ij =b iN ;

[0117] En ij =(b iN -a iN ) / 3;

[0118] He ij =ε iN ;

[0119] Where, ε i1 ε ij and ε iN It is a constant.

[0120] Example 4

[0121] A diagnostic method for power optical transmission networks based on a cloud model evidence cloud droplet mechanism, building upon Example 3, derives a comprehensive cloud model of qualitative indicators based on the aforementioned reliability level cloud model, specifically as follows:

[0122] Obtain the reliability level cloud model C of the k-th expert on the i-th qualitative indicator. ik ;

[0123] The comprehensive cloud model C for calculating the i-th qualitative indicator. i =(Ex j En j He j ):

[0124]

[0125] Among them, Ex i En i He i The integrated cloud model C is respectively i Expectation, entropy, hyperentropy; λ k Let be the weight of the k-th expert, k = 1, 2, 3, ..., t, where t is the number of experts; Ex ik Enik He ik These represent the expected value, entropy, and hyperentropy of the cloud model for the reliability level of the i-th qualitative indicator by the k-th expert.

[0126] Example 5

[0127] A diagnostic method for power optical transmission networks based on the cloud model evidence cloud droplet mechanism, building upon Example 4, obtains the grade support of qualitative indicators based on the comprehensive cloud model, specifically as follows:

[0128] Determine the distance between the integrated cloud model and the cloud models of each reliability level, specifically by generating the integrated cloud model C using a forward cloud generator. i Given T evidence cloud droplets, the reliability level cloud model C of evidence cloud droplet m is calculated using the following formula. j Center (Ex) j Euclidean geometric distance of (,0):

[0129]

[0130] Calculate the basic probability assignment function for each level of evidence cloud droplet m:

[0131]

[0132] Where, m im (A j (C) is a comprehensive cloud model. i The basic probability assignment function of evidence cloud drop m to level j, i.e. the probability that evidence cloud drop m supports level j;

[0133] Calculate the divergence of evidence cloud droplet m:

[0134]

[0135] Among them, KL(m im ||m il ) represents the divergence of evidence cloud drop m, indicating the degree of closeness between evidence cloud drop m and evidence cloud drop l in terms of support at each level. The smaller the value, the higher the degree of closeness.

[0136] Calculate the credibility of evidence cloud droplet m:

[0137]

[0138] The credibility of the evidence cloud droplets was normalized:

[0139]

[0140] The normalized confidence level ω of T evidence cloud drops is obtained. im =[ω i1 ,ω i2,...,ω iT ] T ;

[0141] Based on the normalized confidence level of the evidence cloud droplets, determine the comprehensive probability assignment function of evidence cloud droplet m to level j;

[0142]

[0143] According to the DS evidence theory composition rules, calculate the support of the i-th qualitative indicator for grade j:

[0144]

[0145] in, This refers to the conflict factor in the synthesis formula of evidence theory.

[0146] Example 6

[0147] A diagnostic method for power optical transmission networks based on the cloud model evidence cloud droplet mechanism, in addition to Example 5, specifically calculates the membership degree of qualitative indicators as follows:

[0148]

[0149] in, This is the sum of the correlations between all qualitative indicators and the optimal solution set at level j; This is the sum of the correlations between all qualitative indicators and the worst-case solution set of level j.

[0150] Calculate the membership degree of all qualitative indicators relative to each level:

[0151]

[0152] The overall membership degree of this power optical transmission network to each level is obtained:

[0153] K j (U)=K j1 (U)+K j2 (U).

[0154] When using the maximum membership degree as a qualitative indicator for the diagnostic result, let

[0155] Based on the principle of maximum membership, j0 is taken as the final diagnostic evaluation result.

[0156] The following is a specific embodiment of the present invention.

[0157] Based on relevant data and expert scores, the actual data of each quantitative indicator and the rating of the qualitative indicators for a certain region were obtained, as shown in Tables 2 and 3.

[0158] Table 2 Original values ​​of quantitative indicators

[0159] Indicator Name Fiber optic cable aging rate Average utilization rate of optical core Average configuration of node devices Original value 34.27％ 24.49％ 2.7219

[0160] Table 3 Expert rating results for qualitative indicators

[0161] Expert rating Expert 1 Expert 2 Expert 3 Expert 4 Expert 5 Completeness of spare parts Level 4 Level 3 Level 4 Level 3 Level 3 Scalability of network management system Level 4 Level 3 Level 2 Level 3 Level 3

[0162] According to the method of the present invention, the relationship between the comprehensive cloud model of qualitative indicators and the cloud model of reliability level is shown in Figure 2.

[0163] By combining the weights of the comprehensive indicators and the support of the indicators for each level, the comprehensive support of the regional power backbone communication network for the five reliability levels is obtained, as shown in Table 4.

[0164] Table 4 shows the overall support of this power optical transmission network for five reliability levels.

[0165] Level support Level 1 Level 2 Level 3 Level 4 Level 5 Overall support 0.1294 0.4102 0.7366 0.6671 0.5726

[0166] As shown in Table 4, the final overall network level is Level 3: Medium. This power optical transmission network still has considerable room for development. In subsequent construction, the frequency of optical cable replacement should be increased, and the utilization rate of optical cables should be improved.

[0167] Second aspect

[0168] A diagnostic system for power optical transmission networks based on cloud model evidence cloud droplet mechanism includes a data input module and a calculation module. The data input module inputs the numerical values ​​of indicators for diagnosing the power optical transmission network and the expert's evaluation results on the importance of the indicators. The calculation module calculates and outputs the comprehensive membership degree of the power optical transmission network to each level according to the method described in the first aspect.

[0169] Third aspect

[0170] A storage medium storing a program that, when executed by a computer, outputs the comprehensive membership degree of the power optical transmission network to each level according to the method described in the first aspect.

[0171] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure made using the contents of the present invention specification and drawings, or directly or indirectly applied to other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism, characterized in that, Includes the following steps: Determine the various indicators used for diagnosing power optical transmission networks, including qualitative and quantitative indicators; Determine the weights of the indicators; A reliability level cloud model for qualitative indicators is determined. A comprehensive cloud model for qualitative indicators is obtained based on the reliability level cloud model. The level support degree of qualitative indicators is obtained based on the comprehensive cloud model. The level membership degree of qualitative indicators is calculated, and the maximum membership degree is taken as the diagnostic result of the qualitative indicator. The power optical transmission network is diagnosed based on the results of quantitative and qualitative indicators. The cloud model for determining the reliability level of the qualitative indicators is specifically as follows: The qualitative indicators are divided into N reliability levels; Let C ij =(Ex ij En ij He ij ) represents the reliability level cloud model corresponding to qualitative index i, where C i1 C iN and C ip The corresponding reliability level cloud models are semi-falling cloud, semi-rising cloud, and normal cloud, respectively; p = 2, 3, ..., N-1; The digital characteristics of the reliability level cloud are calculated using the following method: For semi-falling clouds C i1 : Ex ij =a i1 ; In ij =(b i1 -to i1 ) / 3; He ij =ε i1 ; For normal cloud C ip : Ex ij =(b ij +a ij ) / 2; In ij =(b ij -to ij ) / 6; He ij =ε ij ; Where j = p; For half-liter cloud C iN : Ex ij =b iN ; In ij =(b iN -to iN ) / 3; He ij =ε iN ; Where, ε i1 ε ij and ε iN It is a constant; The specific steps for calculating the membership degree of qualitative indicators are as follows: in, This is the sum of the correlations between all qualitative indicators and the optimal solution set at level j; This is the sum of the correlations between all qualitative indicators and the worst-case solution set of level j; Calculate the membership degree of all qualitative indicators relative to each level:

2. The diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism according to claim 1, characterized in that, The qualitative indicators include the completeness of spare parts and the scalability of the network management system, while the quantitative indicators include the aging rate of optical cables, the average utilization rate of optical cores, and the average configuration of node equipment.

3. The diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism according to claim 1, characterized in that, The determination of the indicator weights is specifically as follows: The indicators were compared pairwise using a scaling method, specifically by experts who assessed their importance. An indicator was assigned a rating of 1, 3, 5, 7, and 9 relative to another indicator: equally important, slightly important, significantly important, strongly important, and extremely important, respectively. The results were then presented in the following matrix: in, and Let A and J represent the upper and lower limits of the relative reliability comparison results of the i-th and j-th indicators, respectively. According to the interval number operation method, we have A = [A...]. - A + ], where A + A - It is an upper and lower bound matrix; Calculate A using the characteristic root method for interval numbers. + A - The corresponding largest eigenvalues And the normalized feature w with positive components + w - ; Calculate the positive component coefficients m and negative component coefficients k of matrix A: Where n is the number of indicators; Calculate matrix A + A - The corresponding consistency test index values ​​CR + CR - ,when When the value is less than 0.1, the consistency test is passed, and the interval weight vector of the indicator is obtained as W = [kw - ,mw + ].

4. The diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism according to claim 3, characterized in that, Based on the aforementioned reliability level cloud model, a comprehensive cloud model with qualitative indicators is obtained, specifically as follows: Obtain the reliability level cloud model C of the k-th expert on the i-th qualitative indicator. ik ; The comprehensive cloud model C for calculating the i-th qualitative indicator. i =(Ex j En j He j ): Among them, Ex i En i He i The integrated cloud model C is respectively i Expectation, entropy, hyperentropy; λ k Let be the weight of the k-th expert, k = 1, 2, 3, ..., t, where t is the number of experts; Ex ik En ik He ik These represent the expected value, entropy, and hyperentropy of the cloud model for the reliability level of the i-th qualitative indicator by the k-th expert.

5. The diagnostic method for power optical transmission networks based on cloud model evidence cloud droplet mechanism according to claim 4, characterized in that, The grade support of the qualitative indicators is obtained based on the comprehensive cloud model, specifically as follows: Determine the distance between the integrated cloud model and the cloud models of each reliability level, specifically by generating the integrated cloud model C using a forward cloud generator. i Given T evidence cloud droplets, the reliability level cloud model C of evidence cloud droplet m is calculated using the following formula. j Center (Ex) j Euclidean geometric distance of (,0): Calculate the basic probability assignment function for each level of evidence cloud droplet m: Where, m im (A j (C) is a comprehensive cloud model. i The basic probability assignment function of evidence cloud drop m to level j, i.e. the probability that evidence cloud drop m supports level j; Calculate the divergence of evidence cloud droplet m: Among them, KL(m im ||m il ) represents the divergence of evidence cloud drop m, indicating the degree of closeness between evidence cloud drop m and evidence cloud drop l in terms of support at each level. The smaller the value, the higher the degree of closeness. Calculate the credibility of evidence cloud droplet m: The credibility of the evidence cloud droplets was normalized: The normalized confidence level ω of T evidence cloud drops is obtained. im =[ω i1 ,ω i2 ,...,ω iT ] T ; Based on the normalized confidence level of the evidence cloud droplets, determine the comprehensive probability assignment function of evidence cloud droplet m to level j; According to the DS evidence theory composition rules, calculate the support of the i-th qualitative indicator for grade j: in, This refers to the conflict factor in the synthesis formula of evidence theory.

Images