Direct trust evaluation method and system

Abstract

The invention relates to a direct trust evaluation method and system. PKI authentication overhead in the frequent interaction process of a terminal is reduced by introducing a trust evaluation mechanism. A direct trust evaluation mechanism based on a time decay factor is provided by considering the influence of historical interaction, a weighted time decay factor is introduced to each historical interaction, and historical record data is updated. The Beta distribution is improved to enhance the flexibility and adaptability of a direct trust evaluation model, so that the time trend in historical records is better captured, finally, the dynamic change of historical interaction behaviors is more accurately captured, the influence of expired interaction records is weakened, and the timeliness of trust evaluation of interaction terminals in the same region and the same domain of the electric power intelligent terminal is improved.

Description

Technical Field

[0001] The present invention relates to the field of collaboration of electric power intelligent terminals, and in particular to a direct trust evaluation method and system. Background Art

[0002] With the orderly opening of the power grid and the sharp increase in demand for collaborative operations, the "end-to-end" collaborative capabilities of power smart terminals are constantly being built, which puts forward higher requirements for the efficiency, dynamism and adaptability of terminal identity authentication. At the same time, social capital has participated in the investment and operation of incremental power distribution, distributed energy and other fields, and the ownership and operation rights of power grid assets have been separated. The massive and complex terminal interactions have broken the status quo of the original power terminal zoning and domain division. The single identity authentication mode can no longer meet the actual needs, and it is urgent to establish a hierarchical terminal identity authentication mechanism for the terminal.

[0003] Although the existing terminal trust evaluation in the same area and domain takes historical interaction behaviors into consideration, it does not weight interaction records at different times, which can easily lead to the influence of expired historical interactions, resulting in inaccurate evaluation results. Summary of the invention

[0004] In order to solve the above problems, the purpose of the present invention is to provide a direct trust evaluation method and system, which fully considers the historical interaction behavior of terminals in the same domain, and introduces a time decay factor according to the impact of the interaction timeliness on the current trust evaluation. The weight of expired interaction records is weakened, and the dynamic changes of the terminal trust status are more accurately captured, thereby improving the time correlation of the evaluation results of the interactive terminals in the same area and domain of the power intelligent terminals.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A direct trust assessment method, comprising:

[0007] In the same-domain terminal interaction network, the trust value is defined and calculated by quantifying the task execution performance of the terminal node. At a specific time point t, A set of time series probability data representing tasks in the interaction between terminal nodes;

[0008] Considering the temporal impact of historical interaction records, weights are introduced to ensure that time factors are taken into account when evaluating direct trust;

[0009] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time decay function is constructed; for science terminals, the exponential time decay model is replaced with a power function;

[0010] After the time decay factor is introduced, the terminal's historical interaction records are weighted and updated to reflect the impact of their timeliness on the current trust evaluation. The Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust evaluation model is obtained.

[0011] Trust assessment is performed based on the improved direct trust assessment model.

[0012] Furthermore, in the interaction at time point t, terminal n i and terminal n j The interactive task is represented as a set of time series probability arrays:

[0013]

[0014] in, Indicates the ratio of the number of successfully executed tasks to the total number of task interactions within that time.

[0015] Furthermore, considering the temporal impact of historical interaction records, weights are introduced to ensure that time factors are taken into account when evaluating direct trust. The specific definitions are as follows:

[0016]

[0017] Among them, ε i Indicates time t i The time decay factor of

[0018] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time attenuation function is constructed as follows:

[0019]

[0020] Among them, t1, t2, …, t n are the time demarcation points, δ1,δ2,…,δ n is the decay rate of each interval. As the time boundary points increase point by point, the decay rate decreases point by point.

[0021] Furthermore, for rational terminals, the exponential time decay model is replaced by a power function, as follows:

[0022] ε i =(1+t+t i ) -k

[0023] Among them, k is a parameter that controls the decay rate.

[0024] Furthermore, after the time decay factor is introduced, the historical interaction records of the terminal are weighted and updated to reflect the impact of its timeliness on the current trust evaluation, specifically:

[0025] Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is:

[0026] R={R1,R2,…,R m ,…};

[0027] Each record R i is a triplet R i =(t i ,a i ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable;

[0028] The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, and W(t) is defined as:

[0029] W(t)={R i |t-Δt<t i <t};

[0030] That is, the window W(t) contains all interaction records within Δt time before time t;

[0031] Define interaction record R i =(t i ,a i ,s i )’s trust contribution value c i for:

[0032] c i =ε i ×s i ;

[0033] c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results.

[0034] Calculate the weighted trust evaluation value λ(t) based on the window W(t):

[0035]

[0036] Where |W(t)| represents the size of the window, that is, the number of records in the window;

[0037] When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as:

[0038] W(t new )={R i |t new -Δt<t i <t new};

[0039]

[0040] Finally, the above sliding window and trust evaluation value updates are applied to each historical interaction record for updating, and new historical record data is obtained:

[0041]

[0042] Furthermore, the Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust evaluation model is obtained, as follows:

[0043] Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points;

[0044] Note interactive terminal n i and n j The trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t));

[0045] Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively;

[0046] At time t, a new interaction record R is collected i =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows:

[0047]

[0048] α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

[0049] A direct trust evaluation system includes: a same-domain terminal interaction network unit, a direct trust evaluation unit, specifically:

[0050] The same-domain terminal interaction network unit defines and calculates the trust value by quantifying the task execution performance of the terminal node at a specific time point t. A set of time series probability data representing tasks in the interaction between terminal nodes;

[0051] The direct trust evaluation unit performs trust evaluation based on the improved direct trust evaluation model;

[0052] The improved direct trust evaluation model is constructed as follows:

[0053] Considering the temporal impact of historical interaction records, weights are introduced to ensure that time factors are taken into account when evaluating direct trust;

[0054] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time decay function is constructed; for science terminals, the exponential time decay model is replaced by a power function;

[0055] After introducing the time decay factor, the terminal's historical interaction records are weighted and updated to reflect the impact of their timeliness on the current trust assessment; the Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust assessment model is obtained.

[0056] Furthermore, in the interaction at time point t, terminal n i and terminal n j The interactive task is represented as a set of time series probability arrays:

[0057]

[0058] in, Indicates the ratio of the number of successfully executed tasks to the total number of task interactions within that time.

[0059] Furthermore, the temporal influence of historical interaction records is considered, and weights are introduced to ensure that the time factor is taken into account when evaluating direct trust. The specific definitions are as follows:

[0060]

[0061] Among them, ε i Indicates time t i The time decay factor of

[0062] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time attenuation function is constructed as follows:

[0063]

[0064] Among them, t1, t2, …, t n are the time demarcation points, δ1,δ2,…,δ n is the decay rate of each interval. As the time boundary points increase point by point, the decay rate decreases point by point.

[0065] Furthermore, for the rational terminal, the exponential time decay model is replaced by a power function, as follows:

[0066] ε i =(1+t+t i ) -k

[0067] Among them, k is a parameter that controls the decay rate.

[0068] Furthermore, after the time decay factor is introduced, the historical interaction records of the terminal are weighted and updated to reflect the impact of its timeliness on the current trust evaluation, specifically:

[0069] Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is:

[0070] R={R1,R2,…,R m ,…};

[0071] Each record R i is a triplet R i =(t i ,a i ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable;

[0072] The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, and W(t) is defined as:

[0073] W(t)={R i |t-Δt<t i <t};

[0074] That is, the window W(t) contains all interaction records within Δt time before time t;

[0075] Define interaction record R i =(t i ,a i,s i )’s trust contribution value c i for:

[0076] c i =ε i ×s i ;

[0077] c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results.

[0078] Calculate the weighted trust evaluation value λ(t) based on the window W(t):

[0079]

[0080] Where |W(t)| represents the size of the window, that is, the number of records in the window;

[0081] When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as:

[0082] W(t new )={R i |t new -Δt<t i <t new};

[0083]

[0084] Finally, the above sliding window and trust evaluation value updates are applied to each historical interaction record for updating, and new historical record data is obtained:

[0085]

[0086] Furthermore, the improved Beta distribution can more accurately capture the dynamic changes of historical interaction behaviors, and obtain an improved direct trust evaluation model, which is as follows:

[0087] Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points;

[0088] Note interactive terminal n i and n jThe trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t));

[0089] Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively;

[0090] At time t, a new interaction record R is collected i =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows:

[0091]

[0092] α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

[0093] The present invention has the following beneficial effects:

[0094] The present invention reduces the PKI authentication overhead in the process of frequent terminal interactions by introducing a trust evaluation mechanism. Considering the impact of historical interactions, a direct trust evaluation mechanism based on time decay factor is proposed. A weighted time decay factor is introduced for each historical interaction, and the historical record data is updated. The Beta distribution is improved to enhance the flexibility and adaptability of the direct trust evaluation model, so as to better capture the time trend in the historical records, and finally achieve more accurate capture of the dynamic changes of historical interaction behaviors, weaken the impact of expired interaction records, and improve the timeliness of trust evaluation of interactive terminals in the same area and domain of power smart terminals. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is a flow chart of a method in one embodiment of the present invention. DETAILED DESCRIPTION

[0096] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0097] refer to Figure 1 In this embodiment, a direct trust evaluation method is provided, including:

[0098] In the same-domain terminal interaction network, the trust value is defined and calculated by quantifying the task execution performance of the terminal node. At a specific time point t, A set of time series probability data representing tasks in the interaction between terminal nodes;

[0099] Considering the temporal impact of historical interaction records, weights are introduced to ensure that time factors are taken into account when evaluating direct trust;

[0100] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time decay function is constructed; for science terminals, the exponential time decay model is replaced by a power function;

[0101] After the time decay factor is introduced, the terminal's historical interaction records are weighted and updated to reflect the impact of their timeliness on the current trust evaluation. The Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust evaluation model is obtained.

[0102] Trust assessment is performed based on the improved direct trust assessment model.

[0103] In this embodiment, in the interaction at time point t, terminal n i and terminal n j The interactive task is represented as a set of time series probability arrays:

[0104]

[0105] in, Indicates the ratio of the number of successfully executed tasks to the total number of task interactions within that time.

[0106] In this embodiment, the temporal influence of historical interaction records is considered, and weights are introduced to ensure that the time factor is taken into account when evaluating direct trust. The specific definitions are as follows:

[0107]

[0108] Among them, ε i Indicates time t i The time decay factor of

[0109] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time attenuation function is constructed as follows:

[0110]

[0111] Among them, t1, t2, …, tn are the time demarcation points, δ1,δ2,…,δ n is the decay rate of each interval. As the time boundary points increase point by point, the decay rate decreases point by point.

[0112] In this embodiment, for rational terminals, the exponential time decay model is replaced by a power function, as follows:

[0113] ε i =(1+t+t i ) -k

[0114] Among them, k is a parameter that controls the decay rate.

[0115] In this embodiment, after the time decay factor is introduced, the historical interaction records of the terminal are weighted and updated to reflect the impact of its timeliness on the current trust evaluation, specifically:

[0116] Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is:

[0117] R={R1,R2,…,R m ,…};

[0118] Each record R i is a triplet R i =(t i ,a i ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable;

[0119] The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, and W(t) is defined as:

[0120] W(t)={R i |t-Δt<t i <t};

[0121] That is, the window W(t) contains all interaction records within Δt time before time t;

[0122] Define interaction record R i =(t i ,a i ,s i )’s trust contribution value c i for:

[0123] c i =ε i ×s i ;

[0124] c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results.

[0125] Calculate the weighted trust evaluation value λ(t) based on the window W(t):

[0126]

[0127] Where |W(t)| represents the size of the window, that is, the number of records in the window;

[0128] When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as:

[0129] W(t new )={R i |t new -Δt<t i <t new};

[0130]

[0131] Finally, the above sliding window and trust evaluation value updates are applied to each historical interaction record for updating, and new historical record data is obtained:

[0132]

[0133] In this embodiment, the Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust evaluation model is obtained, which is as follows:

[0134] Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points;

[0135] Note interactive terminal n i and n j The trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t));

[0136] Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively;

[0137] At time t, a new interaction record R is collectedi =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows:

[0138]

[0139] α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

[0140] A direct trust evaluation system includes: a same-domain terminal interaction network unit, a direct trust evaluation unit, specifically:

[0141] The same-domain terminal interaction network unit defines and calculates the trust value by quantifying the task execution performance of the terminal node at a specific time point t. A set of time series probability data representing tasks in the interaction between terminal nodes;

[0142] The direct trust evaluation unit performs trust evaluation based on the improved direct trust evaluation model;

[0143] The improved direct trust evaluation model is constructed as follows:

[0144] Considering the temporal impact of historical interaction records, weights are introduced to ensure that time factors are taken into account when evaluating direct trust;

[0145] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time decay function is constructed; for science terminals, the exponential time decay model is replaced by a power function;

[0146] After introducing the time decay factor, the terminal's historical interaction records are weighted and updated to reflect the impact of their timeliness on the current trust assessment; the Beta distribution is improved to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust assessment model is obtained.

[0147] In this embodiment, in the interaction at time point t, terminal n iand terminal n j The interactive task is represented as a set of time series probability arrays:

[0148]

[0149] in, Indicates the ratio of the number of successfully executed tasks to the total number of task interactions within that time.

[0150] In this embodiment, the temporal influence of the historical interaction records is considered, and weights are introduced to ensure that the time factor is taken into account when evaluating direct trust. The specific definitions are as follows:

[0151]

[0152] Among them, ε i Indicates time t i The time decay factor of

[0153] For control business scenarios with high requirements for fine-grained differentiation of time scales, a segmented time attenuation function is constructed as follows:

[0154]

[0155] Among them, t1, t2, …, t n are the time demarcation points, δ1,δ2,…,δ n is the decay rate of each interval. As the time boundary points increase point by point, the decay rate decreases point by point.

[0156] In this embodiment, for the rational terminal, the exponential time decay model is replaced by a power function, as follows:

[0157] ε i =(1+t+t i ) -k

[0158] Among them, k is a parameter that controls the decay rate.

[0159] In this embodiment, after the time decay factor is introduced, the historical interaction records of the terminal are weighted and updated to reflect the impact of its timeliness on the current trust evaluation, specifically:

[0160] Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is:

[0161] R={R1,R2,…,R m ,…};

[0162] Each record R i is a triplet R i =(t i ,ai ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable;

[0163] The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, and W(t) is defined as:

[0164] W(t)={R i |t-Δt<t i <t};

[0165] That is, the window W(t) contains all interaction records within Δt time before time t;

[0166] Define interaction record R i =(t i ,a i ,s i )’s trust contribution value c i for:

[0167] c i =ε i ×s i ;

[0168] c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results.

[0169] Calculate the weighted trust evaluation value λ(t) based on the window W(t):

[0170]

[0171] Where |W(t)| represents the size of the window, that is, the number of records in the window;

[0172] When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as:

[0173] W(t new )={R i |t new -Δt<t i <t new};

[0174]

[0175] Finally, the above sliding window and trust evaluation value updates are applied to each historical interaction record for updating, and new historical record data is obtained:

[0176]

[0177] In this embodiment, the improved Beta distribution is used to more accurately capture the dynamic changes of historical interaction behaviors, and an improved direct trust evaluation model is obtained, which is as follows:

[0178] Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points;

[0179] Note interactive terminal n i and n j The trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t));

[0180] Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively;

[0181] At time t, a new interaction record R is collected i =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows:

[0182]

[0183] α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

[0184] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A direct trust assessment method, characterized in that: include: In the same-domain terminal interaction network, the trust value is defined and calculated by quantifying the task execution performance of the terminal node; Based on the above definition, a time decay factor is introduced, and the time decay factor is used to ensure that the direct trust evaluation model takes the time factor into account when evaluating direct trust; For control business scenarios, a segmented time decay function is constructed; for management terminals, the exponential time decay model is replaced with a power function; After the time decay factor is introduced, the historical interaction records of the terminal are weighted updated, wherein the weighted update is used to reflect the impact of the timeliness of the interaction records on the current trust evaluation; Improve the Beta distribution and obtain an improved direct trust evaluation model; Trust assessment is performed based on the improved direct trust assessment model.

2. A direct trust assessment method according to claim 1, characterized in that: In the same-domain terminal interaction network, the trust value is defined and calculated by quantifying the task execution performance of the terminal node, as follows: In the interaction at time t, terminal n i and terminal n j The interactive task is represented as a set of time series probability arrays: in, It represents the ratio of the number of successfully executed tasks to the total number of task interactions within the time. This metric is used to describe the success rate of cooperative tasks, that is, the probability of task success.

3. A direct trust assessment method according to claim 2, characterized in that: Considering the temporal influence of historical interaction records, a time decay factor is introduced to ensure that the direct trust evaluation model takes time factors into account when evaluating direct trust. The specific definition is as follows: Among them, ε i Indicates time t i The time decay factor of Indicates time t i The ratio of the number of successfully executed tasks to the total number of task interactions; For control business scenarios, a segmented time attenuation factor function is constructed as follows: Among them, t1, t2, …, t n is the time demarcation point, δ1(t-t1) is the decay rate of the time interval (t-t1), and similarly δ2(t-t2), δ n (tt n ) are the decay rates of the corresponding time intervals. The time boundary points increase point by point, and the decay rates decrease point by point.

4. A direct trust assessment method according to claim 3, characterized in that: For the rational terminal, the exponential time decay model is replaced by a power function, as follows: ε i =(1+t+t i ) -k Among them, k is the parameter that controls the decay rate, and t is the time dividing point.

5. A direct trust assessment method according to claim 4, characterized in that: After the time decay factor is introduced, the historical interaction records of the terminal are weighted updated, specifically: Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is: R={R1,R2,…,R m ,…}; Each record R i is a triplet R i =(t i ,a i ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable; The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, then W(t) is defined as: W(t)={R i |t-Δt<t i <t}; That is, the window W(t) contains all interaction records within Δt time before time t; Define interaction record R i =(t i ,a i ,s i )’s trust contribution value c i for: c i =e i ×s i ; c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results. Calculate the weighted trust evaluation value λ(t) based on the window W(t): Where |W(t)| represents the size of the window, that is, the number of records in the window; When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as: W(t new )={R i |t new -Δt<t i <t new }; Finally, the above sliding window W(t new ) and the trust evaluation value λ(t new ) is applied to each historical interaction record to update it, and new historical record data is obtained 6. A direct trust assessment method according to claim 1, characterized in that: The improved Beta distribution is used to obtain an improved direct trust evaluation model, which is as follows: Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points; Note interactive terminal n i and n j The trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t)); Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively; At time t, a new interaction record R is collected i =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows: α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

7. A direct trust evaluation system, characterized in that: include: The same-domain terminal interaction network unit defines and calculates the trust value by quantifying the task execution performance of the terminal node in the same-domain terminal interaction network; A direct trust assessment unit, which performs trust assessment based on the improved direct trust assessment model; Introducing weights, the weights are used to ensure that the direct trust assessment model takes time factors into account when assessing direct trust; For control business scenarios, construct a segmented time attenuation function; For management terminals, the exponential time decay model is replaced by a power function; After the time decay factor is introduced, the historical interaction records of the terminal are weighted updated, wherein the weighted update is used to reflect the impact of its timeliness on the current trust evaluation; Improve the Beta distribution and obtain an improved direct trust evaluation model; Trust assessment is performed based on the improved direct trust assessment model. A direct trust evaluation system according to claim 7, characterized in that, in the interaction at time point t, terminal n i and terminal n j The interactive task is represented as a set of time series probability arrays: in, It represents the ratio of the number of successfully executed tasks to the total number of task interactions within the time. This metric is used to describe the success rate of cooperative tasks, that is, the probability of task success.

8. A direct trust evaluation system according to claim 7, characterized in that: Considering the temporal influence of historical interaction records, weights are introduced to ensure that the direct trust evaluation model takes time factors into account when evaluating direct trust. The specific definitions are as follows: Among them, ε i Indicates time t i The time decay factor of Indicates time t i The ratio of the number of successfully executed tasks to the total number of task interactions; For control business scenarios, a segmented time attenuation factor function is constructed as follows: Among them, t1, t2, …, t n is the time demarcation point, δ1(t-t1) is the decay rate of the time interval (t-t1), and similarly δ2(t-t2), δ n (tt n ) are the decay rates of the corresponding time intervals. The time boundary points increase point by point, and the decay rates decrease point by point.

9. A direct trust evaluation system according to claim 8, characterized in that: For the rational terminal, the exponential time decay model is replaced by a power function, as follows: ε i =(1+t+t i ) -k Among them, k is the parameter that controls the decay rate, and t is the time dividing point.

10. A direct trust assessment method according to claim 9, characterized in that: After the time decay factor is introduced, the historical interaction records of the terminal are weighted updated, specifically: Introduce a sliding time window algorithm to update the interaction situation. Suppose the interaction record data flow is: R={R1,R2,…,R m ,…}; Each record R i is a triplet R i =(t i ,a i ,s i ) respectively identify the timestamp when the interaction occurs, the type of interaction behavior, and the result of the interaction, where the interaction result s i is a binary variable; The sliding window function W(t) is introduced to represent the window state at time t. The window size is Δt, and W(t) is defined as: W(t)={R i |t-Δt<t i <t}; That is, the window W(t) contains all interaction records within Δt time before time t; Define interaction record R i =(t i ,a i ,s i )’s trust contribution value c i for: c i =e i ×s i ; c i Indicates the record R i The contribution to the trust evaluation at time t is affected by the recorded time and the interaction results. Calculate the weighted trust evaluation value λ(t) based on the window W(t): Where |W(t)| represents the size of the window, that is, the number of records in the window; When a new interaction record R new =(t new ,a new ,s new ) arrives, the updating process of the sliding window and the trust evaluation value is expressed as: W(t new )={R i |t new -Δt<t i <t new }; Finally, the above sliding window W(t new ) and the trust evaluation value λ(t new ) is applied to each historical interaction record to update it, and new historical record data is obtained 11. A direct trust evaluation method according to claim 7, characterized in that: The improved Beta distribution is used to obtain an improved direct trust evaluation model, which is as follows: Introducing the time decay factor into the Beta trust model, time-weighting the historical interaction records, and assigning different weight coefficients to interactions at different time points; Note interactive terminal n i and n j The trust degree is a random variable θ(t) that changes with time. θ(t) obeys the Beta distribution with parameters α(t) and β(t), that is, θ(t)~Beta(α(t),β(t)); Among them, α(t) and β(t) represent the time-weighted cumulative values ​​of the number of successful interactions and the number of failed interactions up to time t, respectively; At time t, a new interaction record R is collected i =(t i ,a i ,s i ), weight the contribution value of the interaction record, when s i = 1, α(t) = α(t i )·ε i +1,β(t)=β(t i )·ε i ; When s i = 0, α(t) = α(t i )·ε i , β(t)=β(t i )·ε i +1, the improved Beta trust model is as follows: α(t) represents the regulator of the direct trust value, represents the sum of successful records, and η serves as an adjustment to prevent the denominator from reaching zero.

Images