Intelligent supply chain member enterprise supervision system based on credibility function

Abstract

The invention discloses an intelligent supply chain member enterprise supervision system based on a credibility function, and the system comprises a member enterprise Agent which is used for collecting reliability state information; the data preprocessing Agent is used for preprocessing the original data provided by the member enterprise Agent; the information fusion Agent is used for fusing the preprocessed data of each member enterprise through a trust degree function to generate a trust degree matrix, and calculating the weight of each data according to the trust degree matrix so as to carry out weighted fusion; the reliability supervision Agent is used for carrying out incentive supervision on the to-be-supervised enterprise based on the fusion result in combination with a price supervision model and a supply reliability supervision model; the man-machine interaction Agent is used for outputting a supervision result; and the management Agent and the coordination Agent are respectively used for coordinating communication and task distribution among the Agents. According to the method, the MAS-based member enterprise reliability supervision model is constructed, the weight of each source data in the fusion process is reasonably distributed, the final expression is obtained, and fusion of multiple pieces of supervision data is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of supervision information fusion of open-pit mines, and in particular to a smart supply chain member enterprise supervision system based on a trust function. Background Art

[0002] Intelligent management of smart mine supply chains leverages advanced technologies such as the Internet of Things (IoT), big data, cloud computing, and artificial intelligence (AI), achieving transparency and efficiency across the entire supply chain. By deploying sensors, RFID tags, and other equipment, mines can collect massive amounts of data from production, logistics, inventory, and other processes in real time. AI algorithms are then used to cleanse and analyze this data, providing precise support for decision-making. For example, blockchain technology is being used to build supply chain traceability systems, ensuring the verifiability of the entire process from mining to processing, enhancing trust between upstream and downstream companies. Furthermore, cloud computing platforms break down traditional information silos and facilitate collaborative operations among supply chain members. For example, the application of driverless mining vehicles and intelligent warehousing systems has significantly improved resource allocation efficiency. These technologies not only optimize production planning and logistics scheduling but also reduce equipment failure rates through predictive maintenance, establishing a new paradigm for data-driven supply chain management.

[0003] An agent (or intelligent agent) in the field of artificial intelligence is an entity that can autonomously perceive its environment, process information, and take actions to achieve a predetermined goal. It can take the form of software (such as a virtual assistant) or a physical device (such as a robot). According to the classic definition by Stuart Russell and Peter Norvig, an agent perceives its environment through sensors and acts on it through actuators.

[0004] A shortcoming of existing technologies is that existing member enterprise regulatory measures face significant technical bottlenecks. Current regulation primarily relies on administrative measures such as fines and price caps. However, information acquisition requires integrating real-time, heterogeneous data from multiple downstream enterprises, significantly increasing the complexity of data processing. For example, ERP and CRM systems across different companies face difficulties in seamless integration due to differing data formats and interface standards, creating "data silos" that require significant resources for cleansing and conversion. Furthermore, multi-source information fusion faces real-time challenges: Supply chain collaboration requires the immediate processing of massive amounts of IoT device data. However, existing models are limited by inefficient data pre-processing (e.g., time-consuming cleansing and normalization) and insufficient correlation mining, making them unable to quickly respond to dynamic changes. Furthermore, privacy and security issues hinder the sharing of sensitive information. For example, enterprise operational data and customer information are vulnerable to leakage when transmitted across systems, further limiting the comprehensiveness and timeliness of regulation. These shortcomings make the existing regulatory system unable to support the continuous optimization of overall supply chain performance. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of the existing technology. To achieve the above purpose, a smart supply chain member enterprise supervision system based on a trust function is adopted to solve the problems raised in the above background technology.

[0006] A smart supply chain member enterprise supervision system based on trust function, the system adopts a structure divided into physical layer, data layer, information layer, fusion layer, supervision layer, and interaction layer, including: Member enterprise agents are deployed at the physical layer to collect reliability status information of the enterprises to be supervised; The data preprocessing agent is deployed in the data layer and is used to preprocess the raw data provided by the member enterprise agents; The information fusion agent is deployed in the fusion layer and is used to fuse the pre-processed data of each member enterprise through the trust function to generate a trust matrix. The weight of each data is calculated according to the trust matrix to perform weighted fusion. The reliability supervision agent is deployed at the supervision layer and is used to provide incentive supervision to regulated enterprises based on the fusion results combined with the price supervision model and the supply reliability supervision model; Human-computer interaction agent, deployed in the interaction layer, is used to output supervision results; and Management Agent and coordination Agent are used to coordinate communication and task allocation among agents respectively.

[0007] As a further solution of the present invention: the trust function is in the form of an exponential function, and the expression is: ; In the formula, Defined to satisfy the fuzzy exponential function form; With The member companies provide the same indicator of the same enterprise to be regulated, and the trust function between the data is used. , establish a trust matrix : ; for For the i-th row element in A larger value of indicates that the information provided by member enterprise i is trusted by most member enterprises; otherwise, it indicates that the information provided by member enterprise i is less likely to be true data.

[0008] As a further solution of the present invention: the trust matrix The weight of each data is determined by solving the eigenvector corresponding to its maximum modulus eigenvalue, and the eigenvector is normalized.

[0009] As a further solution of the present invention: the price regulation model adopts a price cap regulation method, and the calculation formula is: ; Where, is the weighted average price of the base period; The upper limit of the price adjustment to be determined for the next regulatory period; is the product price index during the regulatory period; Z is the exogenous factor adjustment value; X is the efficiency improvement factor of the regulated link, which reflects the relative relationship between input and output.

[0010] As a further solution of the present invention: the indicators of supply reliability supervision include product qualification rate and supply reliability, and the supply reliability is calculated by the average customer interruption time.

[0011] As a further solution of the present invention: the member enterprise Agent is used for independent evaluation and supports the distributed expansion and fault-tolerant operation of the system.

[0012] As a further solution of the present invention: the information fusion agent uses at least one of a fuzzy inference algorithm and a neural network algorithm to perform data anti-interference processing.

[0013] As a further solution of the present invention: the incentive regulation includes dynamically adjusting the price cap and supply reliability threshold to balance the efficiency of member enterprises and service stability.

[0014] As a further solution of the present invention: the data preprocessing includes data cleaning, denoising, and format standardization operations.

[0015] As a further solution of the present invention: the human-computer interaction Agent is used to support the visual display of supervision results and historical data backtracking.

[0016] Compared with the prior art, the present invention has the following technical effects: By adopting the above technical solutions, a reliability supervision model for member enterprises based on MAS was constructed.

[0017] First, addressing the complexity and distribution of regulatory information sources, a trust-based data fusion method is proposed. This method first defines a fuzzy exponential trust function to quantify the degree of trust between two data sets. Furthermore, a trust matrix is used to measure the overall trust level of the data provided by each member enterprise. This allows for a reasonable allocation of weights for each source in the fusion process, resulting in a final expression for the data fusion estimate, thereby enabling the fusion of multiple regulatory data sets.

[0018] Secondly, we proposed an incentive-based regulatory model and approach for member companies that takes into account service reliability. Based on the integration of regulatory information, we adopted a price cap approach for price regulation, while service reliability regulation focused on supply reliability and product qualification rates. This regulatory model was established to incentivize member companies to strike a balance between improving operational efficiency and ensuring service reliability.

[0019] The regulatory model and methods can provide theoretical guidance for adjusting the regulatory approach of supply chain reliability management, improving regulatory efficiency, steadily advancing the regulation of price and service reliability, and establishing a supply chain reliability regulatory system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the structure of the member enterprise supervision system of the disclosed embodiment of this application. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Please refer to Figure 1 In an embodiment of the present invention, a smart supply chain member enterprise supervision system based on a trust function is provided. The system adopts a structure comprising a physical layer, a data layer, an information layer, a fusion layer, a supervision layer, and an interaction layer, including: Member enterprise agents are deployed at the physical layer to collect reliability status information of the enterprises to be supervised; The data preprocessing agent is deployed in the data layer and is used to preprocess the raw data provided by the member enterprise agents; The information fusion agent is deployed in the fusion layer and is used to fuse the pre-processed data of each member enterprise through the trust function to generate a trust matrix. The weight of each data is calculated according to the trust matrix to perform weighted fusion. The reliability supervision agent is deployed at the supervision layer and is used to provide incentive supervision to regulated enterprises based on the fusion results combined with the price supervision model and the supply reliability supervision model; Human-computer interaction agent, deployed in the interaction layer, is used to output supervision results; and Management Agent and coordination Agent are used to coordinate communication and task allocation among agents respectively.

[0023] In the specific implementation, based on the distributed and autonomous system of each member enterprise in the smart supply chain, a member enterprise reliability supervision model based on multi-agent collaboration is constructed, such as Figure 1 The reliability supervision model is implemented by multiple agents interacting and collaborating, and seven types of agents are established: member enterprise agent, data preprocessing agent, information fusion agent, reliability supervision agent, human-computer interaction agent, coordination agent, and management agent.

[0024] The system adopts a layered architecture consisting of a physical layer, a data layer, an information layer, a fusion layer, a supervision layer, and an interaction layer. Within these layers, different numbers of agents are distributed. These agents each assume different roles within the system, completing different tasks and achieving their own distinct functions, thereby ensuring the stable and effective operation of the supply chain intelligent reliability supervision system. The reliability supervision model has the following characteristics: (1) Distributed member enterprise agents are not only conducive to parallel processing of evaluation and improving evaluation efficiency, but also avoid the shortcomings of traditional expert systems such as large knowledge base and inconvenient maintenance.

[0025] (2) Multiple evaluation methods can be integrated. Each member enterprise agent can include different algorithms, and then obtain the final data through information fusion. Multi-agent collaborative evaluation greatly improves the evaluation capability and efficiency. For example, fuzzy reasoning algorithms can achieve fuzzy diagnosis of incomplete information, and neural network algorithms have strong anti-interference capabilities.

[0026] (3) High reliability and easy scalability. Each member enterprise agent has independent evaluation capabilities. When some member enterprise agents exit the system, the system performance will only be reduced, but not completely paralyzed. Properly increasing the number of member enterprise agents can further improve system performance. In addition, it can also form a larger system with other systems.

[0027] Reliability supervision workflow: The main workflow of the intelligent reliability supervision system can be divided into five stages: information collection, information processing, information fusion, reliability supervision, and result output.

[0028] (1) Information collection. Through multiple member enterprise agents distributed in the environment, the reliability status information of the regulated enterprises is collected, generating multiple data sources. Under normal circumstances, the system first determines the regulated object and then searches for member enterprises that have historical transaction records with the regulated enterprise. The management agent sends messages to each member enterprise agent in the form of broadcast through the communication mechanism. Under normal circumstances, the member enterprise agent is in a listening state. After receiving the command from the management agent, it queries the historical database. If there is a historical transaction with the member enterprise to be regulated, the relevant data will be sent to the information preprocessing agent.

[0029] (2) Information processing: The raw data provided by each member enterprise agent needs to be preprocessed by the data preprocessing agent.

[0030] (3) Information fusion. By coordinating, combining, and complementing the information obtained by multiple member companies, the uncertainty and limitations of information from individual member companies can be overcome, thereby obtaining more accurate results than the measurement values from individual member companies, thereby improving the effectiveness of the supervision system. The fused information is then provided to the reliability supervision agent.

[0031] (4) Reliability supervision. After obtaining the supervision information, the reliability supervision agent combines price supervision with supply reliability supervision to conduct incentive supervision on the supervised enterprises to promote them to improve their own reliability.

[0032] (5) Result output. Finally, the result is transmitted to the human-computer interaction agent. At this point, the supervision process is completed.

[0033] In this embodiment, the core concept of decision-level information fusion is that member enterprises make local judgments on the regulated enterprise based on their respective historical transaction databases. These evaluation results are then transmitted via communication channels to the fusion center agent, which then uses specific criteria to comprehensively analyze these local judgments and make a final decision. However, due to various interference factors, local judgments may be uncertain, making the various criteria used in decision-level fusion not necessarily optimal for the system or globally. To address this issue, a trust function is used to enhance the system's confidence.

[0034] Specifically, the trust function is: Assume that multiple member companies participate in the assessment of the same parameter of the company to be regulated. i member companies and j The data measured by the member companies are and .if The higher the authenticity, The higher the trust of other data. quilt The degree of trust, that is, Come and see The degree to which the data is likely to be true, and the degree of trust between the data measured by multiple member companies is called trust.

[0035] In order to further quantify the trust between the feedback information of each member enterprise, a trust function is defined . express quilt Degree of trust. According to the definition of trust, the definition formula of trust is set as: (1) Where, is a continuous decreasing function, and .

[0036] The general fusion method is to give a fusion upper limit ( >0), for ,make (2) if =0, it is considered that i Member companies and j If the member companies do not trust each other. =1, then it is considered that i Member companies fully trust j If a member enterprise is not trusted by other member enterprises, or is trusted by only a few member enterprises, the information provided by that member enterprise will be deleted during information fusion. This processing is not conducive to making objective judgments on the actual situation, and the fusion results are overly influenced by subjective factors.

[0037] According to the above analysis, the trust function can be Defined as an exponential function, let: (3) From the definition of the above formula, we can see that The smaller the value of The larger the data and Mutual trust between the bigger it is; The value of is 0, then =1. On the contrary, When the value of is large, Very small.

[0038] Since the exponential function exist The upper value is monotonically decreasing from 1 to 0, so it satisfies the properties of the trust function in the upper trust definition formula. In practical applications, when The value exceeds the set upper limit When , it can be considered that the two data no longer trust each other. =0. Then the trust function is in the form of an exponential function, and the expression is: ; (4) In the formula, Defined to satisfy the fuzzy exponential function form; With The member companies provide the same indicator of the same enterprise to be regulated, and the trust function between the data is used. , establish a trust matrix : ; (5) for For the i-th row element in A larger value of indicates that the information provided by member enterprise i is trusted by most member enterprises; otherwise, it indicates that the information provided by member enterprise i is less likely to be true data.

[0039] In this embodiment, the trust matrix The weight of each data is determined by solving the eigenvector corresponding to its maximum modulus eigenvalue, and the eigenvector is normalized.

[0040] In this embodiment, the supervision information fusion process use Indicates the information provided by member company i The weight in the fusion process. The value reflects the information of other member companies. The comprehensive trust level can be used right Perform weighted summation to obtain the expression of data fusion: , (6) Where, The weight coefficient Should meet the following requirements: (7) In the trust matrix B, the trust function Only indicates measured data right The degree of trust in the measured data of all member companies in the system cannot reflect the The level of trust, The true extent of Comprehensively reflect.

[0041] Should be integrated with a In the trust system All the information of , so we need to find a set of non-negative numbers , such that: (8) According to formula (5), formula (7) can be rewritten into matrix form as follows: (9) Where, , .

[0042] because , so the trust matrix B is a non-negative matrix, and the symmetric matrix has the largest modulus eigenvalue , such that: (10) Find as well as The corresponding eigenvector A, and the components in A are satisfied (i=1,2,…,n). Substituting equation (10) into equation (9), we get: (11) Formula (11) can be used as a measure of the comprehensive trust level among the information provided by each member enterprise, that is: (12) Taking into account the weight coefficient The conditions in formula (7) should be met. After normalization, we get: (13) Substituting formula (13) into formula (6), the final result of the information fusion of all member enterprises is: (14) In this embodiment, the reliability supervision of member enterprises is carried out, and the supervision indicators are selected as follows: Specifically, by combining price regulation and reliability regulation, targeting the characteristics of member companies in the supply chain, and comprehensively considering both price and supply reliability factors, the input / output volume can be expanded according to needs and conditions.

[0043] Input indicators: product price / yuan, order lead time / day, plan completion rate / %, average customer interruption time / day.

[0044] Output indicators: revenue / yuan, number of customers / piece, inventory supply time / day, maximum production capacity / piece, product qualification rate / %.

[0045] Price supervision of member companies: The price regulation model adopts the price cap regulation method, and the calculation formula is: ; Where, is the weighted average price of the base period; The upper limit of the price adjustment to be determined for the next regulatory period; is the product price index during the regulatory period; Z is the exogenous factor adjustment value; X is the efficiency improvement factor of the regulated link, which reflects the relative relationship between input and output.

[0046] The determination of X is the core of price regulation and should be determined comprehensively based on historical production efficiency data, taking into account future technological progress, and how to stimulate enterprises to reduce costs and improve reliability.

[0047] In this embodiment, the reliability supervision of member enterprises is carried out in addition to price supervision, and service reliability supervision is also introduced. The two indicators of product qualification rate and supply reliability are selected, and the average customer interruption time is calculated based on the supply reliability.

[0048] In this embodiment, the indicators for supply reliability supervision include product qualification rate and supply reliability, and the supply reliability is calculated based on the average customer interruption time.

[0049] In this embodiment, member enterprise agents are used for independent assessment, supporting the distributed expansion and fault-tolerant operation of the system. The information fusion agent uses at least one of a fuzzy inference algorithm and a neural network algorithm to perform data anti-interference processing. The human-computer interaction agent supports the visualization of regulatory results and the review of historical data.

[0050] In this embodiment, incentive regulation includes dynamically adjusting price caps and supply reliability thresholds to balance the efficiency and service stability of member companies.

[0051] In this embodiment, data preprocessing includes data cleaning, denoising, and format standardization operations. Although the embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents, and all such changes are intended to be within the scope of protection of the present invention.

Claims

1. A smart supply chain member enterprise supervision system based on a trust function, wherein the system adopts a structure comprising a physical layer, a data layer, an information layer, a fusion layer, a supervision layer, and an interaction layer, characterized in that: include: Member enterprise agents are deployed at the physical layer to collect reliability status information of the enterprises to be supervised; The data preprocessing agent is deployed in the data layer and is used to preprocess the raw data provided by the member enterprise agents; The information fusion agent is deployed in the fusion layer and is used to fuse the pre-processed data of each member enterprise through the trust function to generate a trust matrix. The weight of each data is calculated according to the trust matrix to perform weighted fusion. The reliability supervision agent is deployed at the supervision layer and is used to provide incentive supervision to regulated enterprises based on the fusion results combined with the price supervision model and the supply reliability supervision model; Human-computer interaction agent, deployed in the interaction layer, is used to output supervision results; as well as Management Agent and coordination Agent are used to coordinate communication and task allocation among agents respectively.

2. According to claim 1, a smart supply chain member enterprise supervision system based on trust function is characterized in that: The trust function is in the form of an exponential function, and the expression is: In the formula, b ij Defined to satisfy the fuzzy exponential function form; Suppose n member companies provide the same indicator of the same company to be regulated, according to the trust function b between the data ij , establish the trust matrix B: For the i-th row element in B, if A larger value of indicates that the information provided by member enterprise i is trusted by most member enterprises; On the contrary, it indicates that the information provided by member enterprise i is less likely to be true data.

3. The smart supply chain member enterprise supervision system based on trust function according to claim 2 is characterized in that: The trust matrix B determines the weight of each data by solving the eigenvector corresponding to its maximum modulus eigenvalue and normalizing the eigenvector.

4. According to claim 1, a smart supply chain member enterprise supervision system based on trust function is characterized in that: The price regulation model adopts the price cap regulation method, and the calculation formula is: P t =P t-1 (1+R PI -X)±Z; Where, P t-1 is the weighted average price of the base period; P t The upper limit of the price adjustment to be determined in the next regulatory period; R PI is the product price index during the regulatory period; Z is the exogenous factor adjustment value; X is the efficiency improvement factor of the regulated link, which reflects the relative relationship between input and output.

5. The smart supply chain member enterprise supervision system based on trust function according to claim 1 is characterized in that: The indicators of supply reliability supervision include product qualification rate and supply reliability, and the supply reliability is calculated based on the average customer interruption time.

6. According to claim 1, a smart supply chain member enterprise supervision system based on trust function is characterized in that: The member enterprise Agent is used for independent evaluation and supports the distributed expansion and fault-tolerant operation of the system.

7. The intelligent supply chain member enterprise supervision system based on trust function according to claim 1 is characterized in that: The information fusion agent uses at least one of a fuzzy inference algorithm and a neural network algorithm to perform data anti-interference processing.

8. The intelligent supply chain member enterprise supervision system based on trust function according to claim 1 is characterized in that: The incentive regulation includes dynamic adjustment of price caps and supply reliability thresholds to balance the efficiency and service stability of member companies.

9. The intelligent supply chain member enterprise supervision system based on trust function according to claim 1 is characterized in that: The data preprocessing includes data cleaning, denoising, and format standardization operations.

10. The intelligent supply chain member enterprise supervision system based on trust function according to claim 1 is characterized in that: The human-computer interaction agent is used to support the visual display of supervision results and historical data backtracking.

Images