Ad hoc network direct trust evaluation method and system based on multi-strategy fusion

Abstract

The invention discloses an ad hoc network direct trust evaluation method based on multi-strategy fusion, and the method comprises the following steps: firstly, solving a resolution coefficient of a grey correlation analysis method through an improved variable coefficient method; secondly, solving the importance of each index by using a grey correlation analysis method of an improved resolution coefficient; then, an optimal order weight calculation table is constructed according to the index importance solved by the grey correlation degree analysis method through an optimal order graph method, so that the weight of each index of the ad hoc network is calculated; and finally, acquiring behavior evidence index data of the ad hoc network equipment according to tools such as a flow monitoring tool, normalizing the behavior evidence index data, and combining the normalized behavior evidence index data with each index weight to obtain a direct trust evaluation value. The invention provides a more accurate direct trust evaluation method for trust evaluation of ad hoc network equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of self-organizing network device trust evaluation, in particular to a self-organizing network direct trust evaluation method and system based on multi-strategy fusion. BACKGROUND

[0002] As a combination of mobile communication and computer network, self-organizing network has high flexibility and powerful communication system functions. With its rapid deployment capability and high reliability support, it has shown wide application potential in emergency communication, monitoring, command and dispatching, etc. It can be flexibly adjusted according to different application scenarios and is suitable for the needs of fixed and mobile platforms such as vehicles and ships.

[0003] However, the dynamic topology and frequent access and exit in an open environment make the trusted interaction between self-organizing network nodes a key problem. Traditional trust evaluation models often ignore the network dynamics and comprehensive mining of behavior data, resulting in insufficient evaluation accuracy. SUMMARY

[0004] The present application discloses a self-organizing network direct trust evaluation method based on multi-strategy fusion, which effectively realizes comprehensive direct trust evaluation of self-organizing network devices and improves the accuracy of device trust evaluation.

[0005] Technical scheme: The self-organizing network direct trust evaluation method based on multi-strategy fusion disclosed by the present application comprises the following steps:

[0006] S1, determine the behavior evidence index of self-organizing network device direct trust evaluation, introduce absolute deviation and smoothing factor to improve the variation coefficient method, and take the variation coefficient result of the improved variation coefficient method as the resolution coefficient of the grey correlation degree analysis method;

[0007] S2, calculate the behavior evidence index correlation degree by using the grey correlation degree analysis method combined with the resolution coefficient;

[0008] S3, construct an optimal sequence weight calculation table according to the behavior evidence index correlation degree by using the optimal sequence diagram method, so as to calculate the weight of each behavior evidence index;

[0009] S4, obtain behavior evidence data by using a flow monitoring tool, normalize the behavior evidence data by using membership, and combine the behavior evidence data with the behavior evidence index weight to obtain a direct trust evaluation value.

[0010] Further, S1 specifically comprises the following steps:

[0011] S11, construct the behavior evidence index of direct trust evaluation of ad hoc network. The behavior evidence index of ad hoc network device is divided into reliability attribute, security attribute and regularity attribute. Among them, the reliability attribute is subdivided into connection establishment success times, data transmission packet loss rate and task processing time, the security attribute is subdivided into unauthorized access times, illegal connection times and denial of service connection times, and the regularity attribute is subdivided into guessing username times, guessing password times and IP transmission delay. In summary, the behavior evidence index of direct trust evaluation of ad hoc network device is: connection establishment success times, data transmission packet loss rate, task processing time, unauthorized access times, illegal connection times, denial of service connection times, guessing username times, guessing password times and IP transmission delay.

[0012] S12, form the original data matrix after determining the sample and the index, and convert the original data matrix into a standardized matrix:

[0013]

[0014] S12, calculate the mean value according to the standardized matrix Change the standard deviation to calculate the absolute deviation:

[0015]

[0016] The absolute deviation is weak in response to extreme data points, and can more truly reflect the dispersion degree of the whole sample, and has stronger robustness. Among them is the mean value of the jth behavior evidence index under the standardized matrix, y ij is an element in the standardized matrix, D j is the absolute deviation of the jth behavior evidence index, the standardized matrix is composed of the behavior evidence index values of the sample, i=1, 2,..., n;

[0017] S13, when calculating the coefficient of variation, a smoothing factor ε is added as shown below, which enhances the stability in numerical calculation, especially when the mean value approaches 0, which can avoid division by zero error and extreme weighting. V j is the coefficient of variation of the jth behavior evidence index.

[0018]

[0019] S14, normalize the coefficient of variation of each index to obtain the importance result of each index, and take it as the resolution coefficient of the corresponding index. The coefficient of variation result of the behavior evidence index solved by the improved coefficient of variation method is taken as the resolution coefficient of the grey correlation degree analysis method, which avoids subjective selection and also makes the grey correlation degree analysis method can better distinguish the contribution of different indexes, and will not be disturbed by extreme data and lose stability. As shown below:

[0020]

[0021] wherein, p i is the discrimination coefficient of the ith behavior evidence index, V i is the variation coefficient of the ith behavior evidence index, m is the total number of the behavior evidence indexes, V j is the variation coefficient of the jth behavior evidence index.

[0022] Further, S2 comprises the following steps:

[0023] S21, determining the evaluation index system to collect relevant data, and constructing a matrix as follows, wherein m is the number of the behavior evidence indexes, and n is the column number.

[0024]

[0025] S22, performing dimensionless processing on the matrix and determining the optimal value reference data column of the indexes; the reference data column should be an ideal comparison standard, which can be composed of the optimal value or the worst value of each index, or other reference values can be selected according to the evaluation purpose. There are mainly two methods for dimensionless processing: initial value method and mean value method.

[0026] The initial value method is selected for dimensionless processing, as shown in the following formula.

[0027]

[0028] In the formula, i = 1, 2,..., n; k = 1, 2,..., m.

[0029] S23, calculating the absolute value of the difference between each evaluation index data column and the reference data column, as shown in the following formula:

[0030] Δx i (k) = |x0(k) ’ - x i (k) ‘ |

[0031] In the formula, x0(k) ’ is the reference data of the index k in the reference data column;

[0032] S24, judging the correlation between the indexes. The correlation coefficient of the indexes is calculated, as shown in the following formula:

[0033]

[0034] In the formula, p kis the resolution coefficient of index k, which is used to weaken the influence of distortion caused by too large maximum absolute difference and improve the significance of difference between correlation coefficients. The resolution coefficient of the index is solved by using the importance of the index as the resolution coefficient of the corresponding index. m ax is the maximum value of |x0(k)-x i (k)|, Δ m in is the minimum value of |x0(k)-x i (k)|, ζ i (k) is the correlation coefficient of index k.

[0035] S25, calculating the correlation degree of each index:

[0036] Further, S3 includes the following steps:

[0037] S31, constructing the priority weight calculation table according to the index importance results (grey correlation degree) calculated by the grey correlation analysis method combined with the improved coefficient of variation method. The size between indexes is determined according to the grey correlation degree γ k of the index, as shown in the following formula.

[0038]

[0039] Wherein, A ij is the element of the priority weight calculation table, γ i is the grey correlation degree of the behavior evidence index i, γ j is the grey correlation degree of the behavior evidence index j,

[0040] S32, the priority weight calculation table is used to calculate the row sum of each behavior evidence index by the priority order diagram method, and the behavior evidence index score TTL i is obtained, as shown in the following formula.

[0041]

[0042] S33, normalizing the TTL i value, and then the final index weight is obtained as shown in the following formula.

[0043]

[0044] W=(W1,W2,...,W m )

[0045] Wherein, W j is the behavior evidence index weight, and W is the behavior evidence index weight set.

[0046] Further, S4 includes the following steps:

[0047] S41, using a flow monitoring tool to obtain behavior evidence data of the ad hoc network device.

[0048] S42, distinguishing the behavior evidence data according to the greater the better and the smaller the better, the behavior evidence data being initial values of behavior evidence indexes;

[0049] S43, normalizing the behavior evidence data: for the greater the better type data:

[0050]

[0051] for the smaller the better type data:

[0052]

[0053] Wherein, g is the membership degree of the behavior evidence data, representing the membership degree of the fuzzy concept "better", taking values between [0, 1], the greater the value of g, the better; e is the behavior evidence data, Sup(e) and Inf(e) are the upper bound and lower bound of the behavior evidence data value respectively;

[0054] S44, obtaining each behavior evidence data vector: G=(g1, g2,...g n );

[0055] S45, multiplying each processed behavior evidence data and each behavior evidence index weight to obtain the final direct trust evaluation value, such as formula:

[0056]

[0057] Beneficial effects:

[0058] (1) The application introduces absolute deviation and smoothing factor to improve the coefficient of variation method, the response of the deviation to extreme data points is weak, the overall dispersion of the sample can be more truly reflected, and the robustness is stronger. The introduction of the smoothing factor enhances the stability in numerical calculation, and can avoid zero division error and extreme weight.

[0059] (2) In this paper, the resolution coefficient set by human in the grey correlation degree analysis method is solved by the variation coefficient of the improved variation coefficient method, which avoids subjective selection and also makes the grey correlation degree analysis method better distinguish the contribution of different indexes, and improves the reliability of the grey correlation degree method.

[0060] (3) The index importance evaluation calculated by the grey correlation degree analysis method with strong objectivity is introduced into the optimal order diagram method as the index importance input of the optimal order diagram method, the influence of expert subjectivity is eliminated, and the accuracy of direct trust evaluation is improved. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 Flowchart of the present application; DETAILED DESCRIPTION

[0062] The technical solutions of the present application are further described below in combination with the drawings and examples.

[0063] Example 1:

[0064] As Figure 1 shown in a multi-strategy fusion-based direct trust evaluation method for ad hoc networks, comprising the following steps:

[0065] S1, determining the behavior evidence index of the direct trust evaluation of the ad hoc network device, introducing the absolute deviation and the smoothing factor to improve the coefficient of variation method, and taking the coefficient of variation result of the behavior evidence index solved by the improved coefficient of variation method as the resolution coefficient of the grey correlation degree analysis method;

[0066] S2, calculating the correlation degree of the behavior evidence index by using the grey correlation degree analysis method combined with the resolution coefficient;

[0067] S3, constructing an optimal sequence weight calculation table according to the correlation degree of the behavior evidence index by using the optimal sequence diagram method, so as to calculate the weight of each behavior evidence index;

[0068] S4, obtaining the behavior evidence data by using the flow monitoring tool, and obtaining the direct trust evaluation value by combining the normalized behavior evidence data with the behavior evidence index weight by using the membership degree.

[0069] Among them, S1 is specifically divided into the following steps:

[0070] S11, constructing the ad hoc network behavior evidence index. The ad hoc network device behavior evidence index is divided into reliability attribute, security attribute and regularity attribute. Among them, the reliability attribute is subdivided into connection establishment success times, data transmission packet loss rate and task processing time, the security attribute is subdivided into unauthorized access times, illegal connection times and denial of service connection times, and the regularity attribute is subdivided into guessing username times, guessing password times and IP transmission delay. In summary, the ad hoc network device direct trust evaluation behavior evidence index is: connection establishment success times, data transmission packet loss rate, task processing time, unauthorized access times, illegal connection times, denial of service connection times, guessing username times, guessing password times and IP transmission delay.

[0071] S12, forming an original data matrix after determining the sample and the index, and converting the original data matrix into a standardized matrix:

[0072]

[0073] S12, calculating the mean value according to the standardized matrix Change the standard deviation to calculate the absolute deviation:

[0074]

[0075] The absolute deviation is weakly responsive to extreme data points, can more truly reflect the overall dispersion of the sample, and has stronger robustness. Among them is the average of the jth behavior evidence index in the standardized matrix, y ij is an element in the standardized matrix, D j is the absolute deviation of the jth behavior evidence index, the standardized matrix is composed of the behavior evidence index values of the sample, i = 1, 2,..., n;

[0076] S13, when calculating the coefficient of variation, a smoothing factor ε is added as shown below, which enhances the stability in numerical calculation, especially when the mean approaches 0, which can avoid division by zero error and extreme weighting.V j is the coefficient of variation of the jth behavior evidence index.

[0077]

[0078] S14, normalize the coefficient of variation of each index to obtain the importance of each index, and use it as the resolution coefficient of the corresponding index. The coefficient of variation of the behavior evidence index calculated by the improved coefficient of variation method is used as the resolution coefficient of the grey correlation degree analysis method, which avoids subjective selection and also makes the grey correlation degree analysis method better distinguish the contribution of different indicators and not be disturbed by extreme data to lose stability. As shown below:

[0079]

[0080] In this example, 7 groups of samples are randomly selected, and the standardized matrix is constructed and transposed as shown in Table 1. The coefficient of variation is calculated by the improved method, where the smoothing factor is 0.01. The final calculation results are shown in Table 2.

[0081] Table 1 Standardized matrix

[0082]

[0083] Table 2 Calculation results of each index

[0084]

[0085] Among them, S2 is specifically divided into the following steps:

[0086] S21, determine the evaluation index system to collect relevant data and construct a matrix as follows, where m is the number of behavior evidence indexes and n is the number of columns.

[0087]

[0088] S22, dimensionless processing of the matrix and determining the optimal value of the index reference data column; the reference data column should be an ideal comparison standard, which can be composed of the optimal value or the worst value of each index, or other reference values can be selected according to the evaluation purpose. There are mainly two methods for dimensionless processing: initial value method and mean value method.

[0089] The initial value method is selected for dimensionless processing, as shown in the following formula.

[0090]

[0091] In the formula, i = 1, 2, …, n; k = 1, 2, …, m.

[0092] S23, the absolute value of the difference between each evaluation index data column and the reference data column is calculated as shown in the following formula:

[0093] Δx i (k) = |x0(k) ’ -x i (k) ‘ |

[0094] In the formula, x0(k) ’ is the reference data of index k in the reference data column;

[0095] S24, judging the correlation between each index. The correlation coefficient of the index is calculated as shown in the following formula:

[0096]

[0097] In the formula, ρ k is the resolution coefficient of index k, which is used to weaken the distortion caused by the too large maximum absolute difference and improve the significant difference between the correlation coefficients. The resolution coefficient uses the importance of the index solved by the improved coefficient of variation as the resolution coefficient of the corresponding index. Δ m ax is the maximum value of |x0(k)-x i (k)|, Δ m in is the minimum value of |x0(k)-x i (k)|, ζ i (k) is the correlation coefficient of index k.

[0098] S25, calculating the correlation degree of each index:

[0099] In this embodiment, 7 groups of data are randomly selected as samples, and the correlation degrees of each index are calculated by the grey correlation degree analysis method as shown in Table 3. In the grey correlation analysis method, the resolution coefficient is solved by the improved coefficient of variation method, and the optimal sequence is selected as the reference.

[0100] Table 3 index correlation degree results

[0101]

[0102]

[0103] Wherein, S3 is specifically divided into the following steps:

[0104] S31, according to the index importance results (grey correlation degree) calculated by the grey correlation analysis method combined with the improved coefficient of variation method, the priority weight calculation table is constructed. The size between indexes is determined according to the index correlation degree γ k , as shown in the following formula.

[0105]

[0106] Wherein, A ij is the element of the priority weight calculation table, γ i is the grey correlation degree of the behavior evidence index i, γ j is the grey correlation degree of the behavior evidence index j,

[0107] S32, the priority order diagram method calculates the row sum of each behavior evidence index by using the priority weight calculation table, and obtains the behavior evidence index score, wherein TTL i is the behavior evidence index score, as shown in the following formula.

[0108]

[0109] S33, the TTL i value is normalized, and then the final index weight is obtained as shown in the following formula.

[0110]

[0111] W = (W1, W2,..., W m )

[0112] Wherein, W j is the behavior evidence index weight, and W is the behavior evidence index weight set.

[0113] In this example, according to the correlation degree size, the priority weight calculation table is constructed as shown in Table 4:

[0114] Table 4 priority weight calculation table

[0115]

[0116]

[0117] And the final weight is obtained as shown in the following formula:

[0118] W = (0.21, 0.012, 0.16, 0.136, 0.037, 0.099, 0.062, 0.099, 0.185)

[0119] Wherein, S4 is specifically divided into the following steps:

[0120] S41, using flow monitoring tools to obtain the behavior evidence data of the ad hoc network device.

[0121] S42, distinguishing the behavior evidence data according to the greater the better and the smaller the better, the behavior evidence data is the initial value of the behavior evidence index;

[0122] S43, normalizing the behavior evidence data: for the greater the better type data:

[0123]

[0124] For the smaller the better type data:

[0125]

[0126] Wherein, g is the membership degree of the behavior evidence data, representing the membership degree of the fuzzy concept "better", taking value between [0, 1], the greater the value of g, the better; e is the behavior evidence data, Sup(e), Inf(e) are the upper bound and lower bound of the behavior evidence data value respectively;

[0127] S44, obtaining each behavior evidence data vector: G = (g1, g2,...g n );

[0128] S45, multiplying each behavior evidence data after processing with each behavior evidence index weight to obtain the final direct trust evaluation value, such as formula:

[0129]

[0130] In this example, a common handheld device of ad hoc network is selected as the object, and after obtaining its behavior evidence, the normalization processing is carried out according to the optimal membership degree formula, and the result is shown as follows:

[0131] G = (0.62, 0.43, 0.45, 0.61, 0.54, 0.62, 0.73, 0.65, 0.52)

[0132] Combined with the final index weight W, the direct trust value of the ad hoc network device is finally obtained, and the calculation result is shown as follows:

[0133]

[0134] Example 2:

[0135] A multi-strategy fusion-based direct trust evaluation system for ad hoc networks, characterized by comprising a data input module, a traffic monitoring tool, a data processing module, and a data output module;

[0136] The data input module determines the behavior evidence indicators of the direct trust evaluation of the ad hoc network device;

[0137] The traffic monitoring tool obtains behavior evidence data, which is the initial value of the behavior evidence indicators;

[0138] The data processing module introduces absolute deviation and smoothing factor to improve the variation coefficient method, takes the variation coefficient result of the behavior evidence indicators obtained by the improved variation coefficient method as the resolution coefficient of the grey correlation degree analysis method, calculates the correlation degree of the behavior evidence indicators by using the grey correlation degree analysis method combined with the resolution coefficient, and constructs an optimal sequence weight calculation table according to the correlation degree of the behavior evidence indicators by using the optimal sequence diagram method, so as to calculate the weight of each behavior evidence indicator;

[0139] The data output module obtains the direct trust evaluation value by combining the normalized behavior evidence data with the weight of the behavior evidence indicators using the membership degree.

Claims

1. A multi-strategy fusion based direct trust evaluation method for ad hoc networks, characterized in that, The method comprises the following steps: S1, determining the behavior evidence indicators of the direct trust evaluation of the ad hoc network device, introducing absolute deviation and a smoothing factor to improve the coefficient of variation method, and taking the coefficient of variation result of the behavior evidence indicators solved by the improved coefficient of variation method as the resolution coefficient of the grey correlation degree analysis method; S2, calculating the correlation degree of the behavior evidence indicators by using the grey correlation degree analysis method in combination with the resolution coefficient; S3, constructing an optimal sequence weight calculation table according to the correlation degree of the behavior evidence indicators by using the optimal sequence graph method, so as to calculate the weight of each behavior evidence indicator; S4, obtaining the behavior evidence data by using a flow monitoring tool, normalizing the behavior evidence data by using the membership degree, and combining the behavior evidence data with the weight of the behavior evidence indicators to obtain the direct trust evaluation value.

2. The multi-strategy fusion based direct trust evaluation method for ad hoc networks according to claim 1, wherein, In step S1, the behavior evidence indicators of the direct trust evaluation of the ad hoc network device are divided into reliability attributes, security attributes and regularity attributes; the reliability attributes include the number of successful connection establishment, the data transmission packet loss rate and the task processing time, the security attributes include the number of unauthorized access, the number of illegal connection and the number of denial of service connection, and the regularity attributes include the number of guessed usernames, the number of guessed passwords and the IP transmission delay.

3. The multi-strategy fusion based direct trust evaluation method for ad hoc networks according to claim 1, wherein, In step S1, the absolute deviation is introduced to improve the coefficient of variation method, and specifically, the absolute deviation is introduced to replace the standard deviation, as shown in the following formula: wherein is the mean of the jth behavioral evidence indicator under the standardized matrix, y ij is an element in the standardized matrix, D j is the absolute deviation of the jth behavioral evidence indicator, the standardized matrix being composed of the behavioral evidence indicator values of the samples, i = 1, 2,..., n; A smoothing factor is introduced to improve the coefficient of variation method, as shown in the following formula: where V j is the coefficient of variation of the jth behavioral evidence indicator, and ε is a smoothing factor.

4. The multi-strategy fusion based direct trust evaluation method for ad hoc networks according to claim 1, wherein, The coefficient of variation result of the behavior evidence indicators solved by the improved coefficient of variation method is taken as the resolution coefficient of the grey correlation degree analysis method, as shown in the following formula: wherein, ρ i is the discrimination coefficient of the i-th behavior evidence index, V i is the variation coefficient of the i-th behavior evidence index solved by the improved variation coefficient method, m is the total number of behavior evidence indexes, V j is the variation coefficient of the j-th behavior evidence index.

5. The method according to claim 1, wherein, Step S2 specifically includes: S21, collecting sample data and constructing a matrix as follows, wherein m is the number of behavior evidence indicators, and n is the number of columns, S22, performing dimensionless processing on the matrix and determining the optimal value reference data column of the indicators; the initial value method is selected for dimensionless processing, as shown in the following formula: In the formula, i=1, 2,..., n; k=1, 2,..., m; S23, calculating the absolute value of the difference between each behavior evidence indicator data column and the reference data column, as shown in the following formula: Δx i (k) = |x0(k) ’ -x i (k) ‘ | where x0(k) is the reference data at index k in the reference data series; and ’ is the reference data at index k in the reference data series; and S24, judging the correlation between the indicators and calculating the correlation coefficient of the indicators, as shown in the following formula: wherein ρ k is a resolution coefficient for index k, which is the resolution coefficient obtained in step S1, Δ m is a maximum value of |x i (k) - x m is a minimum value of |x i (k) - x i (k) is a correlation coefficient for index k. S25, calculate the correlation degree of each index:

6. The method according to claim 1, wherein, Step S3 specifically includes: S31, the calculation method of the optimal sequence weight calculation table is as follows: wherein A ij is an element of the optimal weight calculation table, γ i is the grey correlation degree of the behavior evidence indicator i, γ j is the grey correlation degree of the behavior evidence indicator j, S32, the priority diagram method uses the priority weight calculation table to calculate the row sum of each behavioral evidence index, to obtain the behavioral evidence index score, where TTL i is the behavioral evidence index score, S33, for TTL i The values ​​are normalized to obtain the behavioral evidence index weight as shown in the following formula: W = (W1, W2,..., W m ) wherein W j is the behavior evidence indicator weight, and W is a set of behavior evidence indicator weights.

7. The method according to claim 1, wherein the method is characterized by, Step S4 specifically includes: S41, obtaining the behavior evidence data of the ad hoc network device by using a flow monitoring tool; S42, distinguishing the behavior evidence data according to the greater the better and the smaller the better, wherein the behavior evidence data is the initial value of the behavior evidence indicators; S43, normalizing the behavior evidence data: for the greater the better type data: For the smaller the better type data: Wherein g is the membership degree of the behavior evidence data, representing the membership degree of the fuzzy concept "optimal", and the value of g is between 0 and 1, and the greater the value of g, the better; e is the behavior evidence data, Sup(e) and Inf(e) are the upper limit and lower limit of the behavior evidence data value respectively; S44, obtaining each behavior evidence data vector: G = (g1, g2,... g n ). S45, multiplying the processed behavior evidence data and the weight of each behavior evidence indicator to obtain the final direct trust evaluation value, as shown in the formula:

8. A multi-strategy fusion based direct trust evaluation system for ad hoc networks, characterized in that, The method comprises a data input module, a flow monitoring tool, a data processing module and a data output module. The data input module determines the behavior evidence indicators of the direct trust evaluation of the ad hoc network device; The traffic monitoring tool acquires behavior evidence data, which is an initial value of the behavior evidence indicators; The data processing module introduces absolute deviation and a smoothing factor to improve the coefficient of variation method, and uses the coefficient of variation result of the behavior evidence indicators solved by the improved coefficient of variation method as a resolution coefficient of the grey correlation degree analysis method; The grey correlation degree analysis method is used to calculate the correlation degree of the behavior evidence indicators by combining the resolution coefficient, and the optimal order weight calculation table is constructed according to the correlation degree of the behavior evidence indicators by using the optimal order graph method, so as to calculate the weight of each behavior evidence indicator; The data output module obtains the direct trust evaluation value by combining the normalized behavior evidence data with the weight of the behavior evidence indicators.

9. The multi-strategy fusion based direct trust evaluation system for ad hoc networks as claimed in claim 8 wherein, The coefficient of variation method is improved by introducing absolute deviation, and the absolute deviation is used to replace the standard deviation, as shown in the following formula: wherein is the mean of the jth behavioral evidence indicator under the standardized matrix, y ij is an element in the standardized matrix, D j is the absolute deviation of the jth behavioral evidence indicator, the standardized matrix being composed of the behavioral evidence indicator values of the sample, i = 1, 2,..., n; The coefficient of variation method is improved by introducing a smoothing factor, as shown in the following formula: Among them, V j is the coefficient of variation of the jth behavioral evidence indicator, and ε is the smoothing factor.

10. The multi-strategy fusion based direct trust evaluation system for ad hoc networks as claimed in claim 8 wherein, The coefficient of variation result of the behavior evidence indicators solved by the improved coefficient of variation method is used as a resolution coefficient of the grey correlation degree analysis method, as shown in the following formula: wherein, ρ i is the discrimination coefficient of the i-th behavior evidence index, V i is the variation coefficient of the i-th behavior evidence index solved by the improved variation coefficient method, m is the total number of behavior evidence indexes, V j is the variation coefficient of the j-th behavior evidence index.

Images