Adaptive quantitative evaluation method and system for digital object trust management of complex energy main body

Abstract

The invention relates to a self-adaptive quantitative evaluation method for trust management of a digital object of a complex energy main body. The method comprises the following steps: constructing a trust index system of the digital object of the energy main body; carrying out data standardization on trust index data in the constructed index system and calculating a weight; performing Beta distribution trust modeling and Bayesian updating on the basis of the obtained standardized data and the weight to obtain a trust value after dynamic updating; and generating a comprehensive trust value between each pair of energy main bodies by adopting a weighted summation method based on the dynamically updated trust value and the calculated weight. According to the invention, the dynamic trust problem among different energy main bodies can be effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy big data, and relates to a dynamic and reliable mutual trust evaluation method and system for different energy entities, in particular to an adaptive quantitative evaluation method and system for digital object trust management of complex energy entities. Background Art

[0002] Energy big data is an important part of the strategic layout of Digital China. The Cyberspace Administration of China, the National Development and Reform Commission, the Ministry of Industry and Information Technology, and the National Energy Administration put forward in the "Digital Collaborative Transformation Development Action Plan" to build an energy digital ecosystem jointly participated by large power grid enterprises, energy supply enterprises, energy equipment enterprises, energy service companies, Internet enterprises and other entities.

[0003] However, the current "centralized" energy big data platform architecture is difficult to be extended into a digital infrastructure for the entire energy industry. First, the cost of interconnected communication and coordination of energy data is high. The cross-entity energy data circulation involves many departments and a large amount of coordination work. Second, there is a lack of an efficient energy data discovery method and mechanism for physical dispersion. The multi-entity energy data is physically dispersed for storage and the data resource standards are not unified. Traditional centralized data indexing and search technologies are difficult to apply, and distributed data search and discovery technologies need to be explored. Third, there is a lack of an energy data element circulation system design and technical support means. There is a lack of a data registration and circulation process mechanism of "joint participation of multiple entities, cross-border multi-source data fusion, and precise and efficient governance means". Through unified data standards, a consensus in the energy industry is formed, and a multi-entity co-construction and circulation mechanism with equal rights and interests is constructed.

[0004] In response to the above problems, in the applications of the energy Internet and the Internet of Things, various energy entities (such as multi-party heterogeneous energy entities of coal, oil, gas, electricity, etc.) can perform data interaction and functional collaboration through digital objects to achieve the overall goal of energy management and optimization. Digital objects not only virtualize physical resources and devices, but also enable efficient data processing and intelligent control, thereby ensuring the accuracy and real-time nature of energy management. Data plays a crucial role in the energy industry, and its authenticity, integrity, and security determine the reliability of the overall operation of the system. However, the lack of trust between these different entities may lead to problems such as unreliable data, increased transaction risks, and reduced cooperation efficiency.

[0005] Therefore, establishing an effective mutual trust evaluation model with digital objects as the core to quantitatively evaluate the trust relationship of data is the key to ensuring safe, efficient, and transparent cooperation among various entities.

[0006] Existing trust assessment methods mostly rely on static trust scores and fail to fully consider the complex and dynamic interaction relationships among entities in the energy system. The deficiencies are specifically manifested in the following aspects: First, the static scoring method cannot reflect the constantly changing interactions among entities in real time, resulting in a lag in trust values and affecting the reliability of the system. Second, traditional methods lack comprehensive consideration of multi-dimensional trust indicators and fail to fully describe the diverse trust relationships among energy entities. In addition, existing models often lack effective management of uncertainties and are difficult to handle the noise and dynamic changes in interactions. These deficiencies make existing methods unable to meet the evolving complex requirements in the energy industry.

[0007] Therefore, it is necessary to propose a trust assessment method for energy entities based on dynamic interaction analysis to solve the above problems and improve the accuracy and dynamic adaptability of trust management among energy entities.

[0008] After retrieval, no publicly available literature on existing technologies identical or similar to the present invention was found. Summary of the Invention

[0009] The purpose of the present invention is to overcome the deficiencies of the existing technologies and propose an adaptive quantitative assessment method and system for trust management of complex energy entity digital objects, which can effectively solve the dynamic trust problems among different energy entities.

[0010] The present invention realizes the solution of its practical problems by adopting the following technical solutions:

[0011] An adaptive quantitative assessment method for trust management of complex energy entity digital objects includes the following steps:

[0012] Construct a trust index system for energy entity digital objects;

[0013] Standardize the trust index data in the constructed index system and calculate the weights;

[0014] Based on the obtained standardized data and weights, conduct Beta distribution trust modeling and Bayesian update to obtain the dynamically updated trust values;

[0015] Based on the dynamically updated trust values and the calculated weights, use the weighted summation method to generate the comprehensive trust values between each pair of energy entities.

[0016] Moreover, the trust index system of the energy entity digital objects includes:

[0017] 1) Data accuracy (D1): The degree of coincidence between the data and the actual situation.

[0018] 2) Data integrity (D2): Whether the data is missing or tampered with.

[0019] 3) Interaction Success Rate (D3): The ratio of the number of successful historical interactions to the total number of interactions.

[0020] 4) Response Time (D4): The average response time to requests.

[0021] 5) Device Reliability (D5): The stability and failure rate of device operation.

[0022] 6) Transaction Performance Rate (D6): The execution status of transaction contracts.

[0023] 7) Network Communication Quality Index (D7): The stability, latency, and packet loss rate of network transmission

[0024] 8) System Load Index (D8): Reflects the saturation degree of system resource utilization rate and processing capacity

[0025] 9) Environmental Security Level (D9): Evaluates the security protection level and threat status of the system operation environment

[0026] Moreover, the specific method for data standardization and entropy value method to calculate weights for the trust index data in the constructed index system is as follows:

[0027] The specific standardization formula is:

[0028]

[0029] Where x ij represents the value of the i-th subject on the j-th index, and max(x j ) and min(x j ) are the maximum and minimum values of the j-th index respectively.

[0030] The standardized data is used to calculate the information entropy of each trust index to measure its dispersion degree: construct a standardized matrix, and calculate the proportion of the i-th subject under the j-th index:

[0031]

[0032] Where p ij represents the proportion of the standardized value of the i-th subject on the j-th index in all observed values of this index.

[0033] Calculate the information entropy of the j-th index:

[0034]

[0035] When p ij = 0, define p ij lnp ij = 0.

[0036] Calculate the weight w of each index according to the entropy valuej :

[0037]

[0038] Weight w j reflects the influence degree of index j on the comprehensive trust value.

[0039] When the weight changes during an interaction, dynamic weight adjustment is performed:

[0040]

[0041] p factor is the performance factor, and p factor = ω 1 ·hist rate + ω 2 ·resp rate + ω 3 ·res util where hist rate is the historical interaction success rate, which is the ratio of the number of successful interactions to the total number of interactions; resp rate is the real-time response time compliance rate, which is the ratio of the efficient response time to the total response time; res util is the resource utilization efficiency; and ω 1 + ω 2 + ω 3 = 1. In the present invention, they are respectively taken as 0.4, 0.3, 0.3, and the value range of the adjustment coefficient δ is [-0.2, 0.2].

[0042] Moreover, the specific method for performing Beta distribution trust modeling and Bayesian update based on the obtained standardized data and weights is as follows:

[0043] Define Beta distribution parameters: For index j of each subject i, define the number of successes α ij and the number of failures β ij .

[0044] These parameters are based on historical interaction data:

[0045] 1) The number of successful interactions: The number of positive events such as successful data transmission and completed transactions.

[0046] 2) The number of failed interactions: The number of negative events such as failed data transmission and transaction defaults.

[0047] The expectation of the Beta distribution trust value: For index j, the trust value expectation of subject i is:

[0048]

[0049] This expected value represents the credibility of entity i on metric j.

[0050] After each new interaction occurs, the parameters of the Beta distribution are adjusted using the Bayesian update method:

[0051] If the new interaction is successful, then:

[0052]

[0053] If the new interaction fails, then:

[0054]

[0055] The updated expected trust value is:

[0056]

[0057] where λ is the time decay factor: t is the current time, t 0 is the last update time, and ρ is the decay coefficient.

[0058] Through Bayesian update, the trust value can immediately reflect the new interaction situation, thus providing a more dynamic and accurate trust assessment.

[0059] Moreover, the specific method for generating the comprehensive trust value between each pair of energy entities by using the weighted summation method based on the dynamically updated trust value and the calculated weight is:

[0060] Use the weight determined by the entropy method and the expected trust value of the Beta distribution to calculate the comprehensive trust value of each entity:

[0061]

[0062] where f(env j ) represent the environmental impact function and the time decay function respectively:

[0063] f(env j ) = w net *f net + w load *f load + w security *f security

[0064] f net = exp(-λ 1 ·net delay ) * (1 - packet loss )

[0065] f load = 1 / (1 + exp(λ2 *(load rate -threshold)))

[0066] f security = security level / max security_level

[0067] where f net represents the network communication quality, net delay , the network latency time (ms), packet loss is the packet loss rate (%), λ 1 is the network latency sensitivity coefficient (taken as 0.001); f load represents the system load condition, load rate represents the system load rate (%), threshold: the load threshold (recommended value: 80%), λ 2 is the load sensitivity coefficient (taken as 0.1); f security represents the environmental security level, security level is the current security level, max security_level is the highest security level;

[0068] g(time j ) represents the time decay function:

[0069] g(time j ) = exp(-μ * (t current - t last ))

[0070] t current is the current timestamp, t last is the timestamp of the last interaction, μ is the time decay coefficient, and its value range is [0.1, 0.5]. For frequent interaction scenarios: μ = 0.1, for general interaction scenarios: μ = 0.3, and for low-frequency interaction scenarios: μ = 0.5.

[0071] An adaptive quantization evaluation system for trust management of complex energy entity digital objects, comprising:

[0072] A trust index system construction module for energy entity digital objects, which constructs a trust index system for energy entity digital objects based on the requirements for security, accuracy, and reliability in the transfer of digital objects by complex heterogeneous energy entities;

[0073] A standardization and weight calculation module, which standardizes the trust index data in the constructed index system and calculates the weights;

[0074] The trust value calculation module performs Beta distribution trust modeling and Bayesian update based on the obtained standardized data and weights to obtain the dynamically updated trust value.

[0075] The comprehensive trust value generation module generates the comprehensive trust value between each pair of energy entities by using the weighted summation method based on the dynamically updated trust value and the calculated weights.

[0076] The indicators in the trust index system construction module of the energy entity digital object include:

[0077] Data accuracy D1

[0078] Data integrity D2

[0079] Interaction success rate D3

[0080] Response time D4

[0081] Device reliability D5

[0082] Delivery compliance rate D6

[0083] Network communication quality index D7

[0084] System load index D8

[0085] Environmental safety level D9

[0086] A storage medium stores a computer program, which, when executed by a processor, implements the adaptive quantization evaluation method for the trust management of a complex energy entity digital object.

[0087] Advantages and beneficial effects of the present invention:

[0088] The present invention proposes an energy management method based on advanced mathematics and statistical models, especially an inter-trust evaluation model applicable to multi-energy entities in the energy Internet, Internet of Things, and distributed energy systems. Through the combination of the entropy method, Beta distribution, and Bayesian update, this model provides a dynamic and reliable inter-trust evaluation method for different energy entities. The multi-dimensional trust evaluation method proposed by the present invention combines the entropy method, Beta distribution, and Bayesian update, fully considering the complex interaction relationships between entities in the energy system, introducing a real-time monitoring mechanism to be able to capture the changing interaction situations between entities in a timely manner, avoiding the lag of trust values, and ensuring the reliability of the system. At the same time, considering multi-dimensional trust indicators comprehensively, it comprehensively describes the diverse trust relationships between energy entities from multiple perspectives. For example, it includes indicators such as the stability of transaction history, the timeliness of interaction response, and the sustainability of cooperation. Description of the Drawings

[0089] Figure 1It is the architecture diagram of the heterogeneous multi - energy main body digital object mutual trust evaluation model of the present invention;

[0090] Figure 2 It is the schematic diagram of the digital object metadata of the early warning information of the present invention. Detailed implementation manners

[0091] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0092] A digital object is a digital representation of an entity or resource in a physical energy system, including three parts: an identifier, metadata, and a data entity. The identifier is the unique identity marker of the digital object, and it does not change with the storage environment, access environment, or content of the digital object. Metadata represents the descriptive information related to the digital object and applications, businesses, and supports the discovery of required digital objects according to application requirements. Metadata represents the descriptive information related to the digital object and applications, businesses, and supports the discovery of required digital objects according to application requirements.

[0093] Different energy entities have their own digital objects for information interaction and functional collaboration. Examples are as follows:

[0094] 1. Power grid enterprises: digital models of power transmission and distribution equipment, power dispatching data, smart meter data, etc.

[0095] 2. Coal production enterprises: coal production equipment data, coal supply data, transaction records, etc.

[0096] 3. Oil supply enterprises: equipment status data, drilling performance data, maintenance records, etc.

[0097] 4. Natural gas transmission companies: user energy demand data, energy efficiency management systems, energy audit reports, etc.

[0098] 5. Energy Internet enterprises: energy trading platform data, Internet of Things platform data, big data analysis models, etc.

[0099] Considering the characteristics of multi - party heterogeneous energy entities such as coal, oil, gas, and electricity, including complex hierarchical structures, multiple node types, diverse communication methods, large data transmission volumes, and high requirements for security and reliability, all of these increase the difficulty of trust evaluation for energy digital objects. Therefore, the multi - factor trust evaluation system for energy digital objects established by the present invention fully considers the following factors:

[0100] 1) Security: To enhance the security of trust evaluation, various different factors such as communication behavior, transmission rate, and data quality can be incorporated into the trust evaluation system of energy digital objects, reducing the errors and risks caused by inaccurate evaluation of a single factor. At the same time, a decentralized evaluation mode can be designed to disperse the calculation and decision - making of trust evaluation to multiple nodes, avoiding the threats posed by single - point failures and attacks to the evaluation system.

[0101] 2) Accuracy: The accuracy of trust evaluation is the ability of the trust model to accurately predict the trust value of energy digital objects. A high-accuracy evaluation model can better reflect the actual state of energy digital objects. To improve the accuracy of trust evaluation, a multi-factor evaluation model can be established and the credibility of trust parameters can be strengthened.

[0102] 3) Lightweight: The lightweight of trust evaluation is judged by the complexity of the evaluation algorithm and the computational overhead. There are significant differences in the computing capabilities of heterogeneous energy entities, and trust evaluation algorithms with high computational overhead are not applicable to evaluating all energy digital objects. Therefore, to meet the response efficiency of the trust evaluation system and improve the service ability of trust evaluation, selecting core trust parameters and reducing the evaluation dimension are effective means to improve computational efficiency. In addition, improving the effective verification of trust parameters and controlling the trust update frequency are also feasible solutions.

[0103] This model calculates the weights of trust indicators through the entropy method, models the probability of trust values using the Beta distribution, and adjusts the trust values in real time through Bayesian update, so as to realize the comprehensive evaluation and dynamic management of multi-dimensional trust information.

[0104] An adaptive quantitative evaluation method for trust management of digital objects of complex energy entities, as Figure 1 shown, includes the following steps:

[0105] Step 1: Construct a trust indicator system for digital objects of energy entities;

[0106] The trust indicator system is the basis of the evaluation model. By clarifying the trust evaluation indicators of multiple energy entities, such as data accuracy, data integrity, interaction success rate, response time, device reliability, and transaction performance rate, etc., comprehensively consider the impact of different aspects on trust. The trust indicator system provides specific directions and bases for subsequent evaluations, and reflects the credibility of the entity in the interaction process.

[0107] The trust indicator system of the digital object of the energy entity includes:

[0108] The following are the trust indicators for each digital object of the energy entity:

[0109] 1) Data accuracy (D1): The degree of coincidence between the data and the actual situation.

[0110] 2) Data integrity (D2): Whether the data is missing or tampered with.

[0111] 3) Interaction success rate (D3): The ratio of the number of successful historical interactions to the total number of interactions.

[0112] 4) Response time (D4): The average response time to requests.

[0113] 5) Equipment Reliability (D5): The stability and failure rate of equipment operation.

[0114] 6) Transaction Performance Rate (D6): The execution status of transaction contracts.

[0115] 7) Network Communication Quality Index (D7): The stability, latency, and packet loss rate of network transmission

[0116] 8) System Load Index (D8): Reflects the saturation degree of system resource utilization rate and processing capacity

[0117] 9) Environmental Security Level (D9): Evaluates the security protection level and threat status of the system operation environment

[0118] Different energy entities focus on the above different trust indicators, as described below:

[0119] 1) Power grid enterprises: D1, D2, D3, D4, D7, D8

[0120] 2) Coal production enterprises: D3, D5, D6, D8, D9

[0121] 3) Oil supply enterprises: D1, D5, D7, D9

[0122] 4) Natural gas transmission companies: D1, D4, D6, D7, D8

[0123] 5) Energy Internet enterprises: D2, D3, D4, D6, D7, D8, D9

[0124] In this step, the definitions and evaluation criteria of each trust indicator are clarified, aiming to provide unified basic data for subsequent standardization and weight calculation. These definitions will be used as inputs to ensure the consistency and quantifiability of each trust indicator in the model and provide a standardized basis for the calculations in subsequent steps.

[0125] Step 2: Standardize the trust indicator data in the constructed indicator system and calculate the weights;

[0126] The specific method for standardizing the trust indicator data in the constructed indicator system and calculating the weights using the entropy method is as follows:

[0127] In this step, the trust indicator data defined in Step 1 is standardized to eliminate the influence of different units and scales. The standardized data will be input into the entropy method calculation to obtain the importance weights of each trust indicator.

[0128] The specific standardization formula is:

[0129]

[0130] where x ij represents the value of the i-th entity on the j-th metric, and max(x j ) and min(x j ) are the maximum and minimum values of the j-th metric respectively. The standardized data is used to calculate the information entropy of each trust metric to measure its dispersion degree:

[0131] Construct a standardized matrix and calculate the proportion of the i-th entity under the j-th metric:

[0132]

[0133] where p ij represents the proportion of the standardized value of the i-th entity on the j-th metric in all observed values of this metric. Calculate the information entropy of the j-th metric:

[0134]

[0135] When p ij = 0, define p ij lnp ij = 0, where m is the number of evaluating entities.

[0136] Calculate the weight w j of each metric according to the entropy value:

[0137]

[0138] The weight w j reflects the influence degree of metric j on the comprehensive trust value.

[0139] When the weight changes during an interaction, dynamic weight adjustment is performed:

[0140]

[0141] p factor is the performance factor, p factor = ω 1 ·hist rate + ω 2 ·resp rate + ω 3 ·res util , hist rate is the historical interaction success rate, which is the ratio of the number of successful interactions to the total number of interactions; resp rate is the real-time response time compliance rate, which is the ratio of the high-efficiency response time to the total response time; res util is the resource utilization efficiency; and ω 1 + ω 2 + ω 3= 1. In the present invention, they are respectively taken as 0.4, 0.3, and 0.3, and the value range of the coefficient δ is adjusted to [-0.2, 0.2].

[0142] The output of step 2 (normalized data and weights) will be used as the input data for the next step (Beta distribution trust modeling and Bayesian update module) to ensure that the influence of each trust metric is reasonably reflected in the model.

[0143] Step 3: Based on the obtained normalized data and weights, perform Beta distribution trust modeling and Bayesian update to obtain the dynamically updated trust values;

[0144] According to the normalized data and weights in step 2, historical interaction data is used to calculate the initial trust values of each trust metric. The Beta distribution is used to describe the range of variation and uncertainty of these trust values, and the Bayesian update method ensures that the trust values can be dynamically adjusted and updated with each new interaction. In this way, the trust assessment model can reflect the trust changes between energy entities in real time.

[0145] The specific method for performing Beta distribution trust modeling and Bayesian update based on the obtained normalized data and weights is as follows:

[0146] Define the Beta distribution parameters: For the metric j of each entity i, define the number of successes α ij and the number of failures β ij .

[0147] These parameters are based on historical interaction data:

[0148] 1) The number of successful interactions: The number of positive events such as successful data transmission and completed transactions.

[0149] 2) The number of failed interactions: The number of negative events such as failed data transmission and transaction defaults.

[0150] The expectation of the Beta distribution trust value: For the metric j, the expected trust value of entity i is:

[0151]

[0152] This expected value represents the credibility of entity i in metric j.

[0153] After each new interaction occurs, use the Bayesian update method to adjust the parameters of the Beta distribution:

[0154] If the new interaction is successful, then:

[0155]

[0156] If the new interaction fails, then:

[0157]

[0158] The expected updated trust value is:

[0159]

[0160] where λ is the time decay factor: t is the current time, t 0 is the last update time, and ρ is the decay coefficient.

[0161] Through Bayesian update, the trust value can immediately reflect the new interaction situation, thus providing a more dynamic and accurate trust assessment.

[0162] The loop of the Bayesian update and trust comprehensive calculation module: After a new interaction occurs, the Bayesian update module immediately reflects the latest result of the interaction by updating the trust value parameters (such as the number of successes α and the number of failures β), and in real-time feeds the updated trust value back to the trust comprehensive calculation module for re-evaluating the comprehensive trust value between the subjects. After each interaction success or failure, the system will immediately update the corresponding parameters to ensure that the trust value can dynamically reflect the current interaction state. At the same time, the time step of data update can be set to be triggered immediately after each interaction, so as to achieve the real-time nature of trust assessment. The update of the trust value is transmitted through a standardized data interface to ensure seamless linkage between modules and improve the overall response efficiency and reliability of the system.

[0163] Step 4: Based on the dynamically updated trust value in Step 3 and the weights calculated in Step 2, use the method of weighted summation to generate the comprehensive trust value between each pair of energy subjects.

[0164] Based on the dynamically updated trust value in Step 3 and the weights calculated in Step 2, use the method of weighted summation to generate the comprehensive trust value between each pair of energy subjects. The calculation of the comprehensive trust value not only considers the trust situation of each trust index, but also reflects the relative importance of each index, thus forming a comprehensive and accurate trust assessment result. Through this module, we can obtain a value that comprehensively reflects the trust degree between energy subjects, providing a strong basis for the cooperation decision-making between energy subjects.

[0165] The specific method of using the method of weighted summation to generate the comprehensive trust value between each pair of energy subjects based on the dynamically updated trust value in Step 3 and the weights calculated in Step 2 is as follows:

[0166] Use the weight determined by the entropy method and the expected trust value of the Beta distribution to calculate the comprehensive trust value of each subject:

[0167]

[0168] where f(env j ) represent the environmental impact function and the time decay function respectively:

[0169] f(env j ) = w net *f net + w load *f load + w security *f security

[0170] f net = exp(-λ 1 ·net delay ) * (1 - packet loss )

[0171] f load = 1 / (1 + exp(λ 2 * (load rate - threshold)))

[0172] f security = security level / max security_level

[0173] where, f net represents the network communication quality, net delay , the network delay time (ms), packet loss is the packet loss rate (%), λ 1 is the network delay sensitivity coefficient (taken as 0.001); f load represents the system load condition, load rate represents the system load rate (%), threshold: the load threshold (recommended value: 80%), λ 2 is the load sensitivity coefficient (taken as 0.1); f security represents the environmental security level, security level is the current security level, max security_level is the highest security level (such as level 5).

[0174] g(time j ) represents the time decay function:

[0175] g(time j ) = exp(-μ * (t current - t last ))

[0176] t current is the current timestamp, t lastis the timestamp of the last interaction, μ is the time decay coefficient, and its value range is [0.1, 0.5]. For the frequent interaction scenario: μ = 0.1, for the general interaction scenario: μ = 0.3, and for the low-frequency interaction scenario: μ = 0.5.

[0177] This comprehensive trust value reflects the overall credibility of entity i in all trust metrics. The comprehensive trust value reflects the overall credibility among different energy entities in multiple dimensions, helping each entity effectively evaluate the credibility of others during the decision-making and cooperation processes in a complex system.

[0178] Through the calculated comprehensive information value, the present invention establishes a risk warning mechanism, and the risk warning thresholds are set as follows:

[0179] Emergency threshold Critical_threshold = 0.3; Warning threshold Warning_threshold = 0.5; Normal threshold Normal_threshold = 0.7

[0180] As Figure 2 shown in the schematic diagram of the digital object metadata of the warning information, the judgment logic of the warning response mechanism is as follows:

[0181]

[0182] At the same time, trust level classification is required, as shown in the following table:

[0183]

[0184] Each trust level corresponds to different management strategies:

[0185] Premium level: Priority resource allocation, simplified authentication process; High level: Normal resource allocation, standard authentication process; Medium level: Limited resource allocation, enhanced authentication requirements; Low level: Minimal resource allocation, strict authentication monitoring; Risk level: Only maintain basic connection, may limit interaction.

[0186] Among different energy entities, they can evaluate whether they can trust each other by comparing their comprehensive trust values. For example, if an energy entity has a relatively high comprehensive trust value for another energy entity, it indicates that the other party has a relatively high credibility under the consideration of multi-dimensional trust metrics, and thus a certain degree of trust can be placed in the other party for cooperation and other interaction behaviors. Mutual trust is a relative concept and may vary according to different energy entities and application scenarios.

[0187] Generally speaking, if the comprehensive trust value is close to or greater than 0.8, it indicates a high level of mutual trust; if the comprehensive trust value is within a relatively high range (between 0.5 and 0.8), it may indicate a relatively high level of trust between the two parties, and relatively important cooperation can be carried out. If the comprehensive trust value is low (less than 0.5), it indicates a low level of trust, and further observation or measures need to be taken to enhance trust, such as strengthening communication, improving interaction behaviors, etc., in order to conduct more in-depth cooperation.

[0188] To better achieve the connection between modules, clear data interfaces and data formats are defined for each step to ensure the effective transfer of information between modules. Specifically:

[0189] 1. Trust Index System Definition Module: The output data format is structured JSON and CSV formats, which are convenient for other modules to parse and process. These data should include the specific names, value ranges, and units of trust indicators to ensure the consistency of data during transmission.

[0190] 2. Data Standardization Module: The input data is required to be the original data output by the trust index system definition module. The standardization module should support multiple input formats and perform dimensionless and normalization processing on the data. The output data should be transmitted in a unified standardized format to ensure that each trust indicator can be compared under the same dimension.

[0191] 3. Bayesian Update Module: The input parameters include the standardized data of each trust indicator and historical interaction data (such as the number of successes α and the number of failures β). The output parameter is the updated trust value, and the data format should include the current trust value and its uncertainty (such as Beta distribution parameters) of each trust indicator.

[0192] 4. Trust Comprehensive Calculation Module: This module receives the trust value data output by the Bayesian update module and the weights of each trust indicator, and finally outputs the comprehensive trust value. The interface should clearly define data fields, trust indicator names, weights, comprehensive trust values, etc., to ensure seamless docking between modules.

[0193] Through the definition of standardized interfaces and clear data formats, it is ensured that modules can be seamlessly connected, reducing errors and system instability caused by data format mismatches, and further improving the overall reliability and maintainability of the system.

[0194] An adaptive quantitative evaluation system for the trust management of complex energy entity digital objects, including:

[0195] A trust index system construction module for energy entity digital objects, which constructs a trust index system for energy entity digital objects based on the requirements for security, accuracy, and reliability during the transfer of digital objects by complex heterogeneous energy entities;

[0196] A standardization and weight calculation module standardizes the trust index data in the constructed index system and calculates the weights.

[0197] A trust value calculation module performs Beta distribution trust modeling and Bayesian update based on the obtained standardized data and weights to obtain the dynamically updated trust value.

[0198] A comprehensive trust value generation module generates the comprehensive trust value between each pair of energy entities by using the method of weighted summation based on the dynamically updated trust value and the calculated weights.

[0199] The indicators in the trust index system construction module of the energy entity digital object include:

[0200] Data accuracy D1

[0201] Data integrity D2

[0202] Interaction success rate D3

[0203] Response time D4

[0204] Device reliability D5

[0205] Delivery performance rate D6

[0206] Network communication quality index D7

[0207] System load index D8

[0208] Environmental safety level D9

[0209] A storage medium stores a computer program, which implements the adaptive quantization evaluation method for the trust management of complex energy entity digital objects when executed by a processor.

[0210] Example 1:

[0211] The specific implementation manner of the present invention is as follows, taking the mutual trust modeling between a power grid enterprise and an energy supply enterprise as an example:

[0212] 1. Scenario description:

[0213] A certain power grid enterprise needs to evaluate the credibility of a new energy supply enterprise to decide whether to establish a long-term cooperative relationship with it.

[0214] Initial conditions are as follows:

[0215] 1) Evaluation period: 3 months

[0216] 2) Interaction frequency: 10 data exchanges per day on average

[0217] 3) Total number of interactions: approximately 900 times

[0218] 4) Data types: power load data, equipment status data, transaction data, etc.

[0219] 2. Collection of trust indicators

[0220] Collect the following data according to the indicators concerned by power grid enterprises:

[0221] 1) Data accuracy (D1): Number of successful times (α1): 850; Number of failed times (β1): 50; Original accuracy rate: 94.4%

[0222] 2) Data integrity (D2): Number of successful times (α2): 870; Number of failed times (β2): 30; Completeness rate: 96.7%

[0223] 3) Interactive success rate (D3): Number of successful times (α3): 855; Number of failed times (β3): 45; Success rate: 95%

[0224] 4) Response time (D4): Number of successful times (α4): 840; Number of failed times (β4): 60; Compliance rate: 93.3%

[0225] 5) Network communication quality (D7): Average delay: 45ms; Packet loss rate: 0.1%; Network quality score: 0.98

[0226] 6) System load (D8): Average load rate: 65%; Peak load rate: 85%; Load score: 0.92

[0227] 3. Calculation of data standardization

[0228] A. Calculate the expected value E[T ij of each indicator. Without considering the attenuation factor:

[0229] E[T 1 = 850 / (850 + 50) = 0.944

[0230] E[T 2 = 870 / (870 + 30) = 0.967

[0231] E[T 3 = 855 / (855 + 45) = 0.950

[0232] E[T 4 = 840 / (840 + 60) = 0.933

[0233] B. Calculation of environmental impact function

[0234] Impact of network communication:

[0235] f net= exp(-0.001 * 45) * (1 - 0.001) = 0.956

[0236] System load impact:

[0237] f load = 1 / (1 + exp(0.1 * (65 - 80))) = 0.920

[0238] Comprehensive environmental impact:

[0239] f(env) = 0.4 * 0.956 + 0.3 * 0.920 + 0.3 * 0.95 = 0.944C. Time decay function (assuming the most recent interaction was 2 hours ago)

[0240] g(time) = exp(-0.3 * 2) = 0.549

[0241] 4. Entropy method for calculating weights

[0242] A. Construct the initial judgment matrix

[0243] Original data (using 4 observation data of 4 interaction subjects as an example):

[0244] Subject 1: [0.944, 0.967, 0.950, 0.933]

[0245] Subject 2: [0.923, 0.945, 0.938, 0.917]

[0246] Subject 3: [0.956, 0.972, 0.963, 0.942]

[0247] Subject 4: [0.912, 0.958, 0.927, 0.925]

[0248] B. Data standardization

[0249] 1) Calculate the proportion p of the i-th subject under the j-th index ij :

[0250] The first observation of Subject 1: p 11 = 0.944 / (0.944 + 0.923 + 0.956 + 0.912) = 0.253The second observation of Subject 1: p 12 = 0.967 / (0.967 + 0.945 + 0.972 + 0.958) = 0.251And so on to calculate all p ij , obtaining the standardized matrix P in 2)

[0251] 2) The complete standardized matrix P:

[0252]

[0253] C. Calculate the information entropy:

[0254] 1) Calculate the information entropy e of the j-th index j :

[0255]

[0256] Similarly, e 2 =-k∑(p 2j ×ln p 2j ) = 0.997, e 3 = 0.998, e 4 = 0.999.

[0257] D. Calculate the weights:

[0258] 1) Calculate the coefficient of variation:

[0259] d 1 = 1 - e 1 = 0.001, d 2 = 0.003, d 3 = 0.002, d 4 = 0.001

[0260] 2) Calculate the weights of each index:

[0261] w 1 = d 1 / ∑d j = 0.083

[0262] Similarly, w 2 = 0.250, w 3 = 0.167, w 4 = 0.083.

[0263] Assume that in the latest interaction, the historical interaction success rate = 0.95, the real-time response time compliance rate = 0.93, and the resource utilization efficiency = 0.92. Then the performance factor is calculated as:

[0264] p factor = 0.4×0.95 + 0.3×0.93 + 0.3×0.92 = 0.935

[0265] The weights after dynamic adjustment:

[0266] w 1 ' = w 1 ×(1 + 0.1×0.935) = 0.083×1.0935 = 0.091

[0267] w 2 ' = w 2×(1 + 0.1×0.935) = 0.250×1.0935 = 0.273

[0268] w 3 ' = 0.0907, w 4 ' = 0.091

[0269] 5. Comprehensive Trust Value Calculation

[0270] S = (0.091×0.944 + 0.273×0.0967 + 0.0.273×0.950 + 0.091×0.933)

[0271] ×0.944×0.549

[0272] = 0.949×0.944×0.549 = 0.492

[0273] 6. Trust Rating and Response Measures

[0274] According to the calculated comprehensive trust value of 0.492, referring to the trust level classification:

[0275] 0.4 ≤ S i <0.6 => Low Trust Level

[0276] Suggested Measures:

[0277] Implement the "Low Level" management strategy: implement the minimum resource allocation strategy; strengthen the interaction monitoring frequency; require a strict authentication process; restrict access rights to sensitive data

[0278] Improvement Suggestions: optimize the data transmission quality and improve the accuracy; improve the network environment and reduce the response time; increase system redundancy and enhance the reliability; conduct regular performance evaluations and optimizations.

[0279] 7. Real - time Update Example

[0280] Assume that a successful data interaction occurs. Considering only the first four indicators, the updated data accuracy indicator is:

[0281]

[0282] Calculate the new expected trust value:

[0283] E[T 1 new = 0.945,

[0284] Then recalculate the comprehensive trust value using the new expected trust value, where p factorTake it as 0.81 and δ as 0.1, then the new weights are adjusted to: 0.28, 0.25, 0.25, 0.22, and the comprehensive information value is updated to:

[0285] S = (0.28 × 0.945 + 0.25 × 0.967 + 0.25 × 0.949 + 0.22 × 0.933) × 0.944 × 0.549

[0286] = 0.491

[0287] The present invention has the following advantages:

[0288] 1. Dynamic adaptability: The model can reflect new interaction results in a timely manner through Bayesian update, and the trust value is adjusted dynamically accordingly.

[0289] 2. Multi-dimensional comprehensive evaluation: The entropy value method assigns reasonable weights to each trust index, comprehensively considering various factors.

[0290] 3. Uncertainty management: The Beta distribution handles the uncertainty in trust evaluation. The more interactions there are, the more stable the trust value becomes.

[0291] 4. Strong applicability: The model can be applied to different energy entities and various interaction scenarios, with broad application potential.

[0292] The quantitative indicators of the invention effect are as follows:

[0293] The effects of the present invention are reflected in the following aspects:

[0294] 1) Dynamic adaptability: The trust value update response time < 100 ms; the abnormal situation detection accuracy rate > 95%; the adaptive adjustment range of model parameters: ±20%

[0295] 2) Multi-dimensional evaluation accuracy: The comprehensive evaluation accuracy rate reaches over 90%; the misjudgment rate < 5%; the missed judgment rate < 3%

[0296] 3) Uncertainty management: The confidence interval of trust evaluation: 95%; the prediction accuracy improves with the increase of the sample size: after 100 interactions: ±10%, after 1000 interactions: ±5%, after 10000 interactions: ±2%

[0297] 4) System performance: The supported concurrent processing capacity: > 1000 TPS; the average response time: < 50 ms; the system availability: > 99.9%

[0298] The innovation of the present invention lies in:

[0299] 1. Multi - dimensional trust index system: The present invention constructs a multi - dimensional trust index system and defines trust evaluation indexes applicable to different energy entities, including data accuracy, data integrity, interaction success rate, response time, equipment reliability, transaction performance rate, etc. These indexes comprehensively describe the trust relationship between different entities from multiple perspectives and provide a basis for subsequent quantitative evaluation.

[0300] Data standardization and entropy value method module, which is used to standardize the data of different trust indexes and calculate the weights of each trust index based on the entropy value method;

[0301] 2. Calculating weights by data standardization and entropy value method: The weights of each trust index are calculated by using data standardization and entropy value method to eliminate the influence of different units and scales. At the same time, the importance weights of each trust index are dynamically determined according to the entropy value to ensure reasonable weight allocation in the multi - dimensional trust index model;

[0302] 3. Dynamic trust modeling based on Beta distribution and Bayesian update: The Beta distribution is used to model the probability of trust values. Through the Bayesian update method, the parameters of the Beta distribution are adjusted in real - time when each new interaction occurs, so that the trust values can dynamically reflect the interaction situation. This mechanism enhances the dynamic adaptability of the model;

[0303] 4. Quantifying uncertainty with information entropy: The concept of information entropy is introduced to quantify the trust information amount of each trust index. By calculating the entropy value, the certainty and uncertainty of each index are evaluated, providing a basis for the dynamic management and weight allocation of trust values in trust evaluation.

[0304] 5. Comprehensive trust calculation and risk response mechanism: By comprehensively calculating the weighted trust values of each trust index, a comprehensive trust value between energy entities is generated, and different risk response mechanisms are set according to the preset risk thresholds to ensure corresponding management and warning measures are taken at different trust levels, improving the security and reliability of the system.

[0305] 6. Modular data interface design: Modular data interfaces and data formats are defined, including JSON and CSV formats, to ensure seamless data transfer between the trust index system, data standardization, Bayesian update, and comprehensive trust calculation modules, enhancing the scalability and maintainability of the system.

[0306] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation manners. Any other implementation manners obtained by those skilled in the art based on the technical solutions of the present invention also belong to the scope of protection of the present invention.

Claims

1. An adaptive quantitative evaluation method for trust management of complex energy subject digital objects, characterized by: The steps include: Based on the security, accuracy and reliability requirements of complex and heterogeneous energy entities in the circulation of digital objects, a trust index system for digital objects of energy entities is constructed; Standardize the trust indicator data in the constructed indicator system and calculate the weight; Based on the obtained standardized data and weights, Beta distribution trust modeling and Bayesian updating are performed to obtain dynamically updated trust values; Based on the dynamically updated trust value and the calculated weight, a weighted summation method is used to generate the comprehensive trust value between each pair of energy entities.

2. According to claim 1, the adaptive quantitative evaluation method for trust management of complex energy subject digital objects is characterized by: The trust indicator system of the energy subject digital object includes: 1) Data accuracy D1 2) Data integrity D2 3) Interaction success rate D3 4) Response time D4 5) Equipment reliability D5 6) Transaction fulfillment rate D6 7) Network communication quality index D7 8) System load index D8 9) Environmental safety level D9.

3. The adaptive quantitative evaluation method for trust management of complex energy subject digital objects according to claim 1 or 2, characterized in that: The specific method of performing data standardization and entropy value method weight calculation on the trust indicator data in the constructed indicator system is as follows: The specific standardization formula is: where x ij represents the value of the i-th subject on the j-th indicator, max(x j ) and min(x j ) are the maximum and minimum values ​​of the j-th index respectively; Construct a standardized matrix and calculate the weight of the i-th subject under the j-th indicator: Among them, p ij It represents the proportion of the standardized value of the jth indicator of the ith subject to all the observed values ​​of the indicator; Calculate the information entropy of the jth indicator: When p ij = 0, define p ij lnp ij =0; Calculate the weight w of each indicator according to the entropy value j : When the weight changes during an interaction, dynamic weight adjustment is performed: p factor is the performance factor, p factor =ω1·hist rate +ω2·resp rate +ω3·res util , hist rate is the historical interaction success rate; resp rate is the real-time response time compliance rate; res util It is the efficiency of resource utilization.

4. The adaptive quantitative evaluation method for trust management of complex energy subject digital objects according to claim 1 or 2, characterized in that: The specific method of Beta distribution trust modeling and Bayesian updating based on the obtained standardized data and weights is: Define the Beta distribution parameters: For each subject i's indicator j, define the number of successes α ij and the number of failures β ij ; Beta distribution trust value expectation: For indicator j, the trust value expectation of subject i is: After each new interaction occurs, the parameters of the Beta distribution are adjusted using the Bayesian update method: If the new interaction is successful, then: If the new interaction fails, then: The updated trust value is expected to be: Where λ is the time decay factor: t is the current time, t0 is the last update time, and ρ is the decay coefficient.

5. The adaptive quantitative evaluation method for trust management of complex energy subject digital objects according to claim 1 or 2, characterized in that: The specific method of generating the comprehensive trust value between each pair of energy entities by weighted summation based on the dynamically updated trust value and the calculated weight is: The comprehensive trust value of each subject is calculated using the weights determined by the entropy method and the trust value expectation of the Beta distribution: Where f(env j ) represent the environmental impact function and time decay function respectively: f(env j )=w net *f net +w load *f load +w security *f security f net =exp(-λ1·net delay )*(1-packet loss ) f load =1 / (1+exp(λ2*(load rate -threshold))) f security =security level / max security_level Among them, f net Indicates the network communication quality, net delay , network delay time (ms), packet loss is the packet loss rate (%), λ1 is the network delay sensitivity coefficient; f load Indicates the system load. rate represents the system load rate (%), threshold: load threshold, λ2 is the load sensitivity coefficient; f security Indicates the environmental security level, security level is the current security level, max security_level It is the highest security level; g(time j ) represents the time decay function: g(time j )=exp(-μ*(t current -t last )) t current is the current timestamp, t last is the timestamp of the last interaction, μ is the time decay coefficient, and its value range is [0.1, 0.5]. For frequent interaction scenarios: μ = 0.1, for general interaction scenarios: μ = 0.3, and for low-frequency interaction scenarios: μ = 0.

5.

6. An adaptive quantitative evaluation system for trust management of complex energy subject digital objects, characterized by: include: The trust indicator system construction module for digital objects of energy entities builds a trust indicator system for digital objects of energy entities based on the security, accuracy and reliability requirements of complex and heterogeneous energy entities in the circulation of digital objects; The standardization and weight calculation module standardizes the trust indicator data in the constructed indicator system and calculates the weight; The trust value calculation module performs Beta distribution trust modeling and Bayesian updating based on the obtained standardized data and weights to obtain the dynamically updated trust value; The comprehensive trust value generation module generates the comprehensive trust value between each pair of energy entities using a weighted summation method based on the dynamically updated trust value and the calculated weight.

7. The adaptive quantitative evaluation system for trust management of complex energy subject digital objects according to claim 6 is characterized by: The indicators in the trust indicator system construction module of the energy subject digital object include: Data Accuracy D1 Data Integrity D2 Interaction success rate D3 Response time D4 Equipment reliability D5 Transaction fulfillment rate D6 Network communication quality index D7 System load indicator D8 Environmental safety level D9.

8. A storage medium having a computer program stored thereon, which, when executed by a processor, implements an adaptive quantitative evaluation method for trust management of complex energy subject digital objects as described in any one of claims 1 to 5.

Images