Supply chain relation management system for enterprise cluster coordination

Abstract

The invention relates to the technical field of enterprise supply chain management, and discloses a supply chain relation management system for enterprise cluster coordination. The log demand analysis and standard formulation module comprehensively analyzes the log recording demand of the supply chain management system and formulates a unified log standard, interface specification and data format; through log demand analysis and standard formulation and cooperation of multi-layer log acquisition and a security module, the system realizes standardization and layered acquisition of full-link log records, so that all business, service and system layer key operations and comprehensive capture of interactive data are ensured; and moreover, measures such as a trusted execution environment, data encryption and multi-layer identity verification are introduced into the acquisition end, so that the security and integrity of the data in the acquisition, transmission and storage processes are effectively guaranteed. Meanwhile, the log data is subjected to preprocessing, duplicate removal and abnormity monitoring, abnormal conditions in the collection process can be found in real time, and a solid data foundation is laid for subsequent accurate analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise supply chain management, and in particular to a supply chain relationship management system coordinated by enterprise clusters. Background Art

[0002] A supply chain management system coordinated by enterprise clusters uses an information-based, collaborative management system to optimize and coordinate the supply chain relationships among multiple enterprises, particularly those within the same cluster or industry chain. This system aims to help enterprises within the cluster achieve resource sharing, information flow, and mutually beneficial cooperation, thereby improving the overall efficiency and responsiveness of the supply chain.

[0003] As disclosed in the publication number CN113077267A, the invention discloses a supply chain relationship management system for enterprise cluster coordination, which belongs to the technical field of enterprise supply chain management and includes a fixed-point layout module, a numerical positioning module, a positioning adjustment module and a relationship analysis module; the fixed-point layout module is used to graphically layout the enterprise supply chain relationship, the numerical positioning module is used to position each component on the enterprise supply chain on the graphical layout of the fixed-point layout module, the positioning adjustment module is used to adjust the positioning position of each component on the supply chain on the graphical layout, and the relationship analysis module is used to analyze the correlation between the initial node and the dynamic node. The invention is provided with a fixed-point layout module and a numerical positioning module, so that the relationship between each component on the supply chain can be calculated and analyzed in a digital manner, making the management of the supply chain relationship more accurate and more objective.

[0004] However, in the existing technology, supply chain relationship management systems usually adopt centralized or distributed log management systems for logging, which mainly record basic information such as operation time, user identity, operation type, system call, etc. In a distributed system environment, the interaction and data transmission between different subsystems are relatively complex, but log records often only record some simple operation information, and key data details are often not recorded in detail. Therefore, these logs usually cannot help us deeply analyze data changes or discover potential abnormal behaviors, thereby affecting the efficiency of troubleshooting and system optimization.

[0005] Therefore, those skilled in the art provide a supply chain relationship management system coordinated by enterprise clusters to solve the above-mentioned problems. Summary of the Invention

[0006] In view of the deficiencies of the prior art, the present invention provides a supply chain relationship management system coordinated by enterprise clusters to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a supply chain relationship management system coordinated by enterprise clusters, comprising:

[0008] The log requirement analysis and standard setting module comprehensively analyzes the log recording requirements of the supply chain management system and formulates unified log standards, interface specifications and data formats;

[0009] The multi-layer log collection and security module, based on the log records of the log demand analysis and standard setting module, deploys log collection modules at multiple levels to record key operations and interaction processes, while introducing trusted execution environment security measures at the collection end;

[0010] The real-time intelligent monitoring module, based on the data collected by the multi-layer log collection and security modules, dynamically adjusts the granularity of log records according to historical data and real-time data by introducing intelligent algorithms, and monitors the system status and abnormal behavior in real time;

[0011] The distributed log storage module stores the log data collected by each layer in the distributed system after monitoring by the real-time intelligent monitoring module;

[0012] The full-link log big data analysis module uses a big data analysis platform and intelligent algorithms to monitor and analyze the full-link log data in real time based on the data stored in the distributed log storage module.

[0013] Preferably, the multi-layer log collection and security module includes:

[0014] Multi-layer log collection units are deployed at all levels of the system and are responsible for capturing key operations and interaction data at each layer in real time. Data preprocessing and formatting units clean, deduplicate, and standardize the collected raw log data. Collection anomaly monitoring units monitor anomalies in the log collection process in real time.

[0015] Preferably, the real-time intelligent monitoring module includes:

[0016] The data reception and integration unit is responsible for receiving data from multi-layer log collection modules and performing preliminary integration and format preprocessing on the data; the real-time data analysis unit processes the received data stream at high speed and calculates system performance indicators and status parameters in real time; the intelligent algorithm application unit integrates machine learning and other intelligent algorithms to perform in-depth analysis on the preprocessed data.

[0017] Preferably, the collection method of the multi-layer deployment log collection unit includes:

[0018] Step 1.1: Deploy the log module at the business layer to capture key node data. The adaptive importance weight formula is:

[0019]

[0020] Where N represents the total number of business events captured within the specified time window.

[0021] α i represents the basic importance coefficient of each event i,

[0022] Δt i It is expressed as the time difference between event i and the current moment, Indicates nonlinear attenuation of time, λ represents the time attenuation coefficient, F i It represents the frequency of occurrence of event i within the set time window, and is expressed by the logarithmic function ln(1+F i ) for smoothing;

[0023] Step 1.2 uses a dynamic adjustment algorithm to track cross-service calls at the service layer. The dynamic granularity adjustment formula for service layer log collection is:

[0024]

[0025] Among them, W 服务 Represents the weight of log collection, R i represents the request volume of the i-th service, ln(1+R i ) indicates that the impact of the request volume on log records increases logarithmically, T i represents the response time of the i-th service, T max Indicates the maximum value of service response time. It is used to measure the relationship between response time and maximum value, θ i Indicates the systemic importance index associated with the service;

[0026] Step 1.3 uses a dynamic weight calculation formula at the system level to collect data from the operating system, database, and network layers. The dynamic weight calculation formula is:

[0027]

[0028] Where M represents the total number of system events captured within the monitoring time window,

[0029] α k Indicates the basic importance coefficient of event k, ΔR k represents the magnitude of the change in resource usage caused by event k, Indicates nonlinear attenuation of the change amplitude, μ is the sensitivity adjustment parameter, U k Indicates the instantaneous resource utilization when the event occurs, U max Indicates the preset upper limit of resource utilization, which is smoothly mapped by the tanh function.

[0030] Preferably, the acquisition anomaly monitoring unit monitoring method captures persistent abnormal changes by accumulating small deviations:

[0031] S t =max(0,S t-1 +(X t -μ-k)),

[0032] Among them, X t represents the current acquisition value, μ represents the target mean, k represents the drift parameter, S t Indicates the cumulative deviation value at time t;

[0033] When S t When the preset threshold h is exceeded, it is considered an abnormality.

[0034] Preferably, the trusted execution environment includes a hardware security module and a secure computing platform. The hardware security module is used as follows:

[0035] Step 2.1: Use the HSM's built-in hardware random source and physical unclonable function to extract and generate the initial key K0;

[0036] Step 2.2 adopts a dynamic update mechanism combining chaotic mapping and time factor, and its algorithm formula is:

[0037]

[0038] Among them, K t Indicates the new key generated at the current moment, K t-1 Indicates the key at the previous moment, PUF seed represents the unique hardware seed extracted by the PUF,

[0039] x t Generated by chaotic mapping, the calculation formula is: x t+1 =μx t (1-x t ),

[0040] The chaos parameter μ ranges from 3.57 to 4.0, x t ×2 32 Indicates converting the chaotic output into a 32-bit integer, T represents the current timestamp, Represents a bitwise exclusive OR operation, and SHA3-256 represents a nonlinear mixing of the above items to generate a 256-bit output key.

[0041] Preferably, the method for using the secure computing platform includes:

[0042] Step 3.1 Use the MPC protocol to distribute the key generation task to multiple independent participants, and provide data to jointly generate the final key K t , and its algorithm formula is:

[0043] K t =MPC(Share1,Share2,…,Share n ),

[0044] Where Share represents the key shard provided by the i-th participant, and n represents the number of participants;

[0045] Step 3.2 performs differential privacy perturbation on sensitive data before encryption to further protect the privacy of log data. The algorithm formula is:

[0046] Among them, D t Represents the original log data, represents Gaussian noise with mean 0 and standard deviation σ;

[0047] Step 3.3: Use the key K generated by MPC t For the disturbed log data D t ′ For encryption, use a symmetric encryption algorithm:

[0048] L t =AES(K t ,D t ′ ),

[0049] Among them, the key K t The recommended length is 256 bits, and the AES encryption mode is GCM.

[0050] Preferably, the method for data integration and preprocessing by the data receiving and integration unit includes:

[0051] Step 4.1 uses DTW to align the small time differences between different time series. The algorithm formula is:

[0052] Let w be the fixed window size,

[0053]

[0054] Where A=(a1,a2,…,a n ) and B=(b1,b2,…,b n ) represents two time series, and W represents the alignment path that satisfies the boundary conditions;

[0055] Step 4.2 uses MinHash and LSH to detect and remove duplicate log records. The algorithm formula is: Among them, A and B represent the feature sets of two log records.

[0056] Step 4.3 uses a fuzzy matching algorithm to match similar content in different source data. The algorithm formula is:

[0057]

[0058] Among them, f i represents the characteristic value of the i-th log, w i represents the weight of the feature, S fusion represents the final event score after fusion.

[0059] Preferably, the real-time data analysis unit processing method:

[0060] Assume that the observed data x at each time step is t , decomposing it into a low-dimensional representation and a sparse error:

[0061] x t =L t w t +s t ,

[0062] Among them, L t Represents the current low-dimensional subspace basis matrix, w t represents the low-dimensional representation coefficient, s t represents sparse abnormal noise;

[0063] To solve w at the same time t and s t , the following optimization problem is usually constructed:

[0064]

[0065] Among them, λ represents the strength of controlling sparse regularization;

[0066] In solving for w t and s t Then, the subspace basis matrix L is updated using gradient descent t :

[0067]

[0068] η represents the learning rate, Indicates L t Gradient calculation of .

[0069] Preferably, the analysis method of the intelligent algorithm application unit is:

[0070] Use ODMD to find the linear operator A so that the state x at each moment tApproximately mapped to the state x at the next moment t+1 :

[0071]

[0072] Among them, A t represents the linear evolution operator at time t, β represents the forgetting factor, ∈ represents a small positive number, and ||·||2 represents the L2 norm.

[0073] The present invention provides a supply chain relationship management system for enterprise cluster coordination. It has the following beneficial effects:

[0074] 1. Through log demand analysis and standard setting, multi-layer log collection and security module coordination, the present invention achieves the standardization and layered collection of full-link log records. This not only ensures the comprehensive capture of key operations and interaction data at each business, service and system level, but also introduces trusted execution environments, data encryption, multi-layer identity authentication and other measures at the collection end, effectively ensuring the security and integrity of data during collection, transmission and storage. At the same time, log data undergoes preprocessing, deduplication and anomaly monitoring, enabling real-time detection of anomalies during the collection process, laying a solid data foundation for subsequent accurate analysis.

[0075] 2. The real-time intelligent monitoring module of the present invention uses data reception, integration and high-speed real-time data analysis units, as well as intelligent algorithms such as integrated machine learning and online dynamic pattern decomposition (ODMD) to achieve dynamic monitoring and rapid response to system status and abnormal behavior. The system can automatically adjust the granularity of log records based on real-time and historical data, which not only saves storage resources when the system is stable, but also refines the monitoring content when anomalies occur to ensure that problems are captured in time. In conjunction with distributed log storage and full-link big data analysis modules, the entire solution not only has efficient data storage and analysis capabilities, but also can achieve dynamic optimization and scheduling of system resources, thereby improving the overall efficiency of supply chain management and decision-making support capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0077] To help those skilled in the art understand the present invention, the following will provide a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only partial embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0078] The present invention is described in detail below with reference to the accompanying drawings:

[0079] Example:

[0080] Please see the attached Figure 1 , an embodiment of the present invention provides a supply chain relationship management system coordinated by enterprise clusters, including;

[0081] The log requirement analysis and standard setting module comprehensively analyzes the log recording requirements of the supply chain management system and formulates unified log standards, interface specifications and data formats;

[0082] The multi-layer log collection and security module, based on the log records of the log demand analysis and standard setting module, deploys log collection modules at multiple levels to record key operations and interaction processes, while introducing trusted execution environment security measures at the collection end;

[0083] The real-time intelligent monitoring module, based on the data collected by the multi-layer log collection and security modules, dynamically adjusts the granularity of log records according to historical data and real-time data by introducing intelligent algorithms, and monitors the system status and abnormal behavior in real time;

[0084] The distributed log storage module stores the log data collected by each layer in the distributed system after monitoring by the real-time intelligent monitoring module;

[0085] The full-link log big data analysis module uses a big data analysis platform and intelligent algorithms to monitor and analyze the full-link log data in real time based on the data stored in the distributed log storage module.

[0086] The multi-layer log collection and security module includes:

[0087] Multi-layer log collection units are deployed at all levels of the system and are responsible for capturing key operations and interaction data at each layer in real time. Data preprocessing and formatting units clean, deduplicate, and standardize the collected raw log data. Collection anomaly monitoring units monitor anomalies in the log collection process in real time.

[0088] Step S1: The collection method of the multi-layer deployment log collection unit includes:

[0089] Step 1.1: Deploy the log module at the business layer to capture key node data. The adaptive importance weight formula is:

[0090]

[0091] Where N represents the total number of business events captured within the specified time window.

[0092] α i represents the basic importance coefficient of each event i,

[0093] Δti It is expressed as the time difference between event i and the current moment, Indicates nonlinear attenuation of time, λ represents the time attenuation coefficient, F i It represents the frequency of occurrence of event i within the set time window, and is expressed by the logarithmic function ln(1+F i ) for smoothing;

[0094] Step 1.2 uses a dynamic adjustment algorithm to track cross-service calls at the service layer. The dynamic granularity adjustment formula for service layer log collection is:

[0095]

[0096] Among them, W 服务 Represents the weight of log collection, R i represents the request volume of the i-th service, ln(1+R i ) indicates that the impact of the request volume on log records increases logarithmically, T i represents the response time of the i-th service, T max Indicates the maximum value of service response time. It is used to measure the relationship between response time and maximum value, θ i Indicates the systemic importance index associated with the service;

[0097] Step 1.3 uses a dynamic weight calculation formula at the system level to collect data from the operating system, database, and network layers. The dynamic weight calculation formula is:

[0098]

[0099] Where M represents the total number of system events captured within the monitoring time window,

[0100] α k Indicates the basic importance coefficient of event k, ΔR k represents the change in resource usage caused by event k, e -μΔRk Indicates nonlinear attenuation of the change amplitude, μ is the sensitivity adjustment parameter, U k Indicates the instantaneous resource utilization when the event occurs, U max Indicates the preset upper limit of resource utilization, which is smoothly mapped by the tanh function.

[0101] Preferably, the acquisition anomaly monitoring unit monitoring method captures persistent abnormal changes by accumulating small deviations:

[0102] S t =max(0,S t-1 +(X t -μ-k)),

[0103] Among them, X t represents the current acquisition value, μ represents the target mean, k represents the drift parameter, S t Indicates the cumulative deviation value at time t;

[0104] When S t When the preset threshold h is exceeded, it is considered an abnormality.

[0105] Step S2: The trusted execution environment includes a hardware security module and a secure computing platform. The hardware security module is used as follows:

[0106] Step 2.1: Use the HSM's built-in hardware random source and physical unclonable function to extract and generate the initial key K0;

[0107] Step 2.2 adopts a dynamic update mechanism combining chaotic mapping and time factor, and its algorithm formula is:

[0108]

[0109] Among them, K t Indicates the new key generated at the current moment, K t-1 Indicates the key at the previous moment, PUF seed represents the unique hardware seed extracted by the PUF,

[0110] x t Generated by chaotic mapping, the calculation formula is: x t+1 =μx t (1-x t ),

[0111] The chaos parameter μ ranges from 3.57 to 4.0, x t ×2 32 Indicates converting the chaotic output into a 32-bit integer, T represents the current timestamp, Represents a bitwise exclusive OR operation, and SHA3-256 represents a nonlinear mixing of the above items to generate a 256-bit output key.

[0112] Step S3: The method for using the secure computing platform includes:

[0113] Step 3.1 Use the MPC protocol to distribute the key generation task to multiple independent participants, and provide data to jointly generate the final key K t , and its algorithm formula is:

[0114] K t =MPC(Share1,Share2,…,Share n ),

[0115] Where Share represents the key shard provided by the i-th participant, and n represents the number of participants;

[0116] Step 3.2 performs differential privacy perturbation on sensitive data before encryption to further protect the privacy of log data. The algorithm formula is:

[0117] Among them, D t Represents the original log data, represents Gaussian noise with mean 0 and standard deviation σ;

[0118] Step 3.3: Use the key K generated by MPC t For the disturbed log data D t ′ For encryption, use a symmetric encryption algorithm:

[0119] L t =AES(K t ,D t ′ ),

[0120] Among them, the key K t The recommended length is 256 bits, and the AES encryption mode is GCM.

[0121] Specifically, through log demand analysis and standard setting, multi-layer log collection and security module coordination, the system has achieved standardization and layered collection of full-link log records, which not only ensures the comprehensive capture of key operations and interaction data at each business, service and system level, but also introduces trusted execution environment, data encryption, multi-layer identity authentication and other measures at the collection end, effectively ensuring the security and integrity of data during collection, transmission and storage. At the same time, log data is pre-processed, deduplicated and monitored for anomalies, which can detect anomalies in the collection process in real time, laying a solid data foundation for subsequent accurate analysis.

[0122] The real-time intelligent monitoring module includes:

[0123] The data reception and integration unit is responsible for receiving data from multi-layer log collection modules and performing preliminary integration and format preprocessing on the data; the real-time data analysis unit processes the received data stream at high speed and calculates system performance indicators and status parameters in real time; the intelligent algorithm application unit integrates machine learning and other intelligent algorithms to perform in-depth analysis on the preprocessed data.

[0124] Step S4: The method for data integration and preprocessing by the data receiving and integration unit includes:

[0125] Step 4.1 uses DTW to align the small time differences between different time series. The algorithm formula is:

[0126] Let w be the fixed window size,

[0127]

[0128] Where A=(a1,a2,…,a n ) and B=(b1,b2,…,b n ) represents two time series, and W represents the alignment path that satisfies the boundary conditions;

[0129] Step 4.2 uses MinHash and LSH to detect and remove duplicate log records. The algorithm formula is: Among them, A and B represent the feature sets of two log records.

[0130] Step 4.3 uses a fuzzy matching algorithm to match similar content in different source data. The algorithm formula is:

[0131]

[0132] Among them, f i represents the characteristic value of the i-th log, w i represents the weight of the feature, S fusion represents the final event score after fusion.

[0133] Step S5: the real-time data analysis unit processing method:

[0134] Assume that the observed data x in each time period t , decomposing it into a low-dimensional representation and a sparse error:

[0135] x t =L t w t +s t ,

[0136] Among them, L t Represents the current low-dimensional subspace basis matrix, w t represents the low-dimensional representation coefficient, s t represents sparse abnormal noise;

[0137] To solve w at the same time t and s t , the following optimization problem is usually constructed:

[0138]

[0139] Among them, λ represents the strength of controlling sparse regularization;

[0140] In solving for w t and s tThen, the subspace basis matrix L is updated using gradient descent t :

[0141]

[0142] η represents the learning rate, Indicates L t Gradient calculation of .

[0143] Preferably, the analysis method of the intelligent algorithm application unit is:

[0144] Use ODMD to find the linear operator A so that the state x at each moment t Approximately mapped to the state x at the next moment t+1 :

[0145]

[0146] Among them, A t represents the linear evolution operator at time t, β represents the forgetting factor, ∈ represents a small positive number, and ||·||2 represents the L2 norm.

[0147] Specifically, through the real-time intelligent monitoring module, with the help of data reception, integration and high-speed real-time data analysis units, as well as the integration of intelligent algorithms such as machine learning and online dynamic pattern decomposition (ODMD), dynamic monitoring and rapid response to system status and abnormal behavior are achieved. The system can automatically adjust the granularity of log records based on real-time and historical data, which not only saves storage resources when the system is stable, but also refines the monitoring content when anomalies occur to ensure that problems are captured in a timely manner. In conjunction with distributed log storage and full-link big data analysis modules, the entire solution not only has efficient data storage and analysis capabilities, but also can realize dynamic optimization and scheduling of system resources, thereby improving the overall efficiency of supply chain management and decision-making support capabilities.

[0148] While embodiments of the present invention have been shown and described, it will be appreciated 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 invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. Enterprise cluster coordinated supply chain relationship management system, characterized by: include; The log requirement analysis and standard setting module comprehensively analyzes the log recording requirements of the supply chain management system and formulates unified log standards, interface specifications and data formats; The multi-layer log collection and security module, based on the log records of the log demand analysis and standard setting module, deploys log collection modules at multiple levels to record key operations and interaction processes, while introducing trusted execution environment security measures at the collection end; The real-time intelligent monitoring module, based on the data collected by the multi-layer log collection and security modules, dynamically adjusts the granularity of log records according to historical data and real-time data by introducing intelligent algorithms, and monitors the system status and abnormal behavior in real time; The distributed log storage module stores the log data collected by each layer in the distributed system after monitoring by the real-time intelligent monitoring module; The full-link log big data analysis module uses a big data analysis platform and intelligent algorithms to monitor and analyze the full-link log data in real time based on the data stored in the distributed log storage module.

2. The enterprise cluster coordinated supply chain relationship management system according to claim 1, characterized in that: The multi-layer log collection and security module includes: Multi-layer log collection units are deployed at all levels of the system and are responsible for capturing key operations and interaction data at each layer in real time. Data preprocessing and formatting units clean, deduplicate, and standardize the collected raw log data. Collection anomaly monitoring units monitor anomalies in the log collection process in real time.

3. The supply chain relationship management system for enterprise cluster coordination according to claim 1, characterized in that: The real-time intelligent monitoring module includes: The data reception and integration unit is responsible for receiving data from multi-layer log collection modules and performing preliminary integration and format preprocessing on the data; the real-time data analysis unit processes the received data stream at high speed and calculates system performance indicators and status parameters in real time; the intelligent algorithm application unit integrates machine learning and other intelligent algorithms to perform in-depth analysis on the preprocessed data.

4. The enterprise cluster coordinated supply chain relationship management system according to claim 2, characterized in that: The collection method of the multi-layer deployment log collection unit includes: Step 1.1: Deploy the log module at the business layer to capture key node data. The adaptive importance weight formula is: Where N represents the total number of business events captured within the specified time window. α i represents the basic importance coefficient of each event i, Δt i It is expressed as the time difference between event i and the current moment, Indicates nonlinear attenuation of time, λ represents the time attenuation coefficient, F i It represents the frequency of occurrence of event i within the set time window, and is expressed by the logarithmic function ln(1+F i ) for smoothing; Step 1.2 uses a dynamic adjustment algorithm to track cross-service calls at the service layer. The dynamic granularity adjustment formula for service layer log collection is: Among them, W 服务 Represents the weight of log collection, R i represents the request volume of the i-th service, ln(1+R i ) indicates that the impact of the request volume on log records increases logarithmically, T i represents the response time of the i-th service, T max Indicates the maximum value of service response time. It is used to measure the relationship between response time and maximum value, θ i Indicates the systemic importance index associated with the service; Step 1.3 uses a dynamic weight calculation formula at the system level to collect data from the operating system, database, and network layers. The dynamic weight calculation formula is: Where M represents the total number of system events captured within the monitoring time window, α k Indicates the basic importance coefficient of event k, ΔR k represents the change in resource usage caused by event k, e -μΔRk Indicates nonlinear attenuation of the change amplitude, μ is the sensitivity adjustment parameter, U k Indicates the instantaneous resource utilization when the event occurs, U max Indicates the preset upper limit of resource utilization, which is smoothly mapped by the tanh function.

5. The supply chain relationship management system for enterprise cluster coordination according to claim 4, characterized in that: The acquisition anomaly monitoring unit monitoring method captures continuous abnormal changes by accumulating small deviations: S t =max(0,S t-1 +(X t -μ-k)), Among them, X t represents the current acquisition value, μ represents the target mean, k represents the drift parameter, S t Indicates the cumulative deviation value at time t; When S t When the preset threshold h is exceeded, it is considered an abnormality.

6. The enterprise cluster coordinated supply chain relationship management system according to claim 5, characterized in that: The trusted execution environment includes a hardware security module and a secure computing platform. The hardware security module is used as follows: Step 2.1: Use the HSM's built-in hardware random source and physical unclonable function to extract and generate the initial key K0; Step 2.2 adopts a dynamic update mechanism combining chaotic mapping and time factor, and its algorithm formula is: Among them, K t Indicates the new key generated at the current moment, K t-1 Indicates the key at the previous moment, PUF seed represents the unique hardware seed extracted by the PUF, x t Generated by chaotic mapping, the calculation formula is: x t+1 =μx t (1-x t ), The chaos parameter μ ranges from 3.57 to 4.0, x t ×2 32 Indicates converting the chaotic output into a 32-bit integer, T represents the current timestamp, Represents a bitwise exclusive OR operation, and SHA3-256 represents a nonlinear mixing of the above items to generate a 256-bit output key.

7. The supply chain relationship management system for enterprise cluster coordination according to claim 6, characterized in that: The method for using the secure computing platform includes: Step 3.1 Use the MPC protocol to distribute the key generation task to multiple independent participants, and provide data to jointly generate the final key K t , and its algorithm formula is: K t =MPC(Share1,Share2,…,Share n ), Where Share represents the key shard provided by the i-th participant, and n represents the number of participants; Step 3.2 performs differential privacy perturbation on sensitive data before encryption to further protect the privacy of log data. The algorithm formula is: Among them, D t Represents the original log data, represents Gaussian noise with mean 0 and standard deviation σ; Step 3.3: Use the key K generated by MPC t For the disturbed log data D t ′ For encryption, use a symmetric encryption algorithm: L t =AES(K t ,D t ′ ), Among them, the key K t The recommended length is 256 bits, and the AES encryption mode is GCM.

8. The supply chain relationship management system for enterprise cluster coordination according to claim 3, characterized in that: The method for data integration and preprocessing by the data receiving and integration unit includes: Step 4.1 uses DTW to align the small time differences between different time series. The algorithm formula is: Let w be the fixed window size, Where A=(a1,a2,…,a n ) and B=(b1,b2,…,b n ) represents two time series, and W represents the alignment path that satisfies the boundary conditions; Step 4.2 uses MinHash and LSH to detect and remove duplicate log records. The algorithm formula is: Among them, A and B represent the feature sets of two log records. Step 4.3 uses a fuzzy matching algorithm to match similar content in different source data. The algorithm formula is: Among them, f i represents the characteristic value of the i-th log, w i represents the weight of the feature, S fusion represents the final event score after fusion.

9. The enterprise cluster coordinated supply chain relationship management system according to claim 8, characterized in that: The real-time data analysis unit processing method: Assume that the observed data x in each time period t , decomposing it into a low-dimensional representation and a sparse error: x t =L t w t +s t , Among them, L t Represents the current low-dimensional subspace basis matrix, w t represents the low-dimensional representation coefficient, s t represents sparse abnormal noise; To solve w at the same time t and s t , the following optimization problem is usually constructed: Among them, λ represents the strength of controlling sparse regularization; In solving for w t and s t Then, the subspace basis matrix L is updated using gradient descent t : η represents the learning rate, Indicates L t Gradient calculation of .

10. The enterprise cluster coordinated supply chain relationship management system according to claim 9, characterized in that: Analysis method of the intelligent algorithm application unit: Use ODMD to find the linear operator A so that the state x at each moment t Approximately mapped to the state x at the next moment t+1 : Among them, A t represents the linear evolution operator at time t, β represents the forgetting factor, ∈ represents a small positive number, and ||·||2 represents the L2 norm.

Images