A multi-dimensional data-based supply chain anomaly intelligent identification method and system

Abstract

The application discloses a kind of based on multi-dimension data's supply chain exception intelligent identification method and system, applied to supply chain exception identification technical field, by obtaining the order information and logistics information of each order in order management system;Based on the time stamp corresponding to each order information and logistics information of each order, obtain the delay duration, according to delay duration, obtain frequency distribution and high frequency deviation amplitude;According to frequency distribution and high frequency deviation amplitude are analyzed, obtain state field conflict point set;According to state field conflict point set, obtain the adjustment operation information of order of existence conflict, adjustment operation information is analyzed, obtain diffusion degree, according to diffusion degree, obtain the root cause distribution of order processing conflict;According to root cause distribution, the fusion parameters of each order are adjusted and analyzed, obtain delay influence path chart, based on delay influence path chart, obtain exception reason identification result, improve the accuracy of supply chain exception identification.

Description

Technical Field

[0001] This invention relates to the field of supply chain anomaly identification technology, and in particular to a method and system for intelligent identification of supply chain anomalies based on multi-dimensional data. Background Technology

[0002] Power grid supply chain management is crucial for the development of power grid business, and its efficiency and accuracy directly impact enterprise operations and market competitiveness. With the deepening of digital transformation, supply chain systems need to integrate multi-source data to support real-time decision-making. However, existing methods often reveal significant shortcomings when handling multi-source heterogeneous data. Many solutions rely too heavily on data integration from a single system, ignoring dynamic conflicts in cross-system data interactions, or lack in-depth analysis of semantic relationships during data fusion, resulting in low accuracy in locating the root causes of anomalies. Summary of the Invention

[0003] To address the aforementioned technical problems, embodiments of the present invention provide a method and system for intelligent identification of supply chain anomalies based on multi-dimensional data, thereby solving the problem that existing technologies cannot accurately identify the causes of supply chain anomalies.

[0004] A first aspect of this invention provides a method for intelligent identification of supply chain anomalies based on multi-dimensional data, comprising: Retrieve order information and logistics information for each order in the order management system; Based on the timestamps corresponding to the order information and logistics information of each order, the delay duration is obtained. Based on the delay duration, the inventory allocation anomaly characteristics are obtained, including frequency distribution and high-frequency deviation range. The cross-system interface fault logs were analyzed based on frequency distribution and high-frequency deviation range to obtain a set of conflict points in the status fields. Based on the set of conflict points in the status field, obtain the adjustment operation information of conflicting orders, analyze the adjustment operation information to obtain the degree of diffusion, and analyze the interaction logs based on the degree of diffusion to obtain the root cause distribution of order processing conflicts. The fusion parameters of each order are adjusted and analyzed based on the root cause distribution to obtain the delay impact path diagram. Based on the delay impact path diagram, the abnormal cause identification results are obtained.

[0005] In one possible implementation of the first aspect, inventory allocation anomaly characteristics are obtained based on the delay duration, including: Determine if the delay duration is greater than a preset threshold. If the delay duration is greater than the preset threshold, mark the order as having a data synchronization error. Analyze the frequency of orders that experience data synchronization anomalies in different time periods to obtain the anomaly trigger frequency; Obtain inventory change logs for time periods where the frequency of abnormal triggering exceeds a preset frequency. Extract and analyze the inventory change logs to obtain inventory allocation anomaly characteristics, including frequency distribution and high-frequency deviation range.

[0006] In one possible implementation of the first aspect, the inventory change log is extracted and analyzed to obtain inventory allocation anomaly characteristics, including: Extract inventory change logs to obtain inventory change information, and calculate the correlation coefficient between inventory update lag time and the number of order allocation errors based on the inventory change information; Based on the correlation coefficient, abnormal records where the current inventory quantity does not match the inventory quantity occupied by the order are obtained, and the absolute value of the deviation for each abnormal record is calculated. The distribution density of the absolute value of the statistical deviation is used to obtain the frequency of anomalies in each deviation interval. The distribution ratio of the frequency of anomalies in different deviation intervals is calculated. Based on the distribution ratio, the range of high-frequency deviations in inventory allocation errors is determined, and the abnormal characteristics of inventory allocation are obtained.

[0007] In one possible implementation of the first aspect, the cross-system interface fault log is analyzed based on frequency distribution and high-frequency deviation range to obtain a set of status field conflict points, including: Based on the frequency distribution and high-frequency deviation range, the delivery information of the order to be evaluated is obtained; By comparing and analyzing delivery information, the time points of delivery trajectory deviation and the geographical scope of impact can be obtained; By analyzing the time points of delivery trajectory deviations and the geographical impact range, a set of conflict points in status fields can be obtained from the cross-system interface fault logs.

[0008] In one possible implementation of the first aspect, comparative analysis of delivery information is performed to obtain the time points and geographical impact ranges of delivery trajectory deviations, including: The spatial distance is calculated by comparing the transportation coordinates in the delivery information with the warehouse location of the order to be evaluated. The delivery information includes the transportation coordinates, the vehicle number, and the delivery time. If the spatial distance is greater than the preset delivery radius threshold, the location matching is determined to be abnormal, and an abnormal result is obtained. Based on the abnormal result, the vehicle's predetermined path coordinates, actual driving path coordinates, and navigation path data are extracted from the logistics tracking platform. Based on the coordinate difference between the predetermined path and the actual path, the trajectory deviation value is obtained. Based on the trajectory deviation value, the time point of the delivery trajectory deviation and the geographical impact range are obtained.

[0009] In one possible implementation of the first aspect, the delivery trajectory deviation time point and geographical impact range are obtained based on the trajectory deviation value, including: Anomaly trajectory pattern analysis is performed based on trajectory deviation values ​​to obtain spatial distribution characteristics and deviation degree characteristics of trajectory deviation. Spatial distribution features are extracted to obtain multiple timestamps corresponding to each coordinate point; Calculate the difference between any two timestamps to obtain multiple misalignment magnitudes. If the misalignment magnitude is greater than a preset time deviation threshold, the timestamp misalignment is determined to be abnormal. Calculate the distance between the current vehicle position when the timestamp misalignment occurs and the vehicle position collected by the GPS device when the timestamp misalignment occurs, and obtain the distance difference. If the distance difference exceeds the preset location error range, it is judged as a location matching anomaly, and the time point of trajectory deviation and geographical impact range are determined.

[0010] In one possible implementation of the first aspect, cross-system interface fault logs are analyzed based on the delivery trajectory deviation time points and geographical impact range to obtain a set of status field conflict points, including: Obtain order change status information, filter the order change status information based on the trajectory deviation time point, and obtain multiple order processing events; Based on the timestamp of the order processing event and the time point of the delivery trajectory deviation, each order processing event is judged to obtain the record of the processing conflict event and the record of the status update omission event; Based on the conflict event handling records and status update omission event records, the change frequency and omission count of each order during the conflict period are obtained; Based on the frequency of changes and the number of omissions, a set of conflict points in the status fields is obtained.

[0011] In one possible implementation of the first aspect, a set of status field conflict points is obtained based on the frequency of changes and the number of omissions, including: Based on the frequency of changes and the number of omissions, the distribution characteristics of conflict events in the time and business dimensions are calculated to obtain the type classification and quantity distribution of conflict events; Based on the type classification and quantity distribution of conflict events, determine the conflict comparison distribution results; Based on the conflict comparison distribution results, interface data transmission information is extracted from the cross-interface communication logs, and the interface data transmission information is analyzed to obtain a set of conflict points in the status field.

[0012] In one possible implementation of the first aspect, the adjustment operation information is analyzed to obtain the degree of diffusion, including: If there is a conflict, the order number in the conflict point set of the status field is matched with the order number in the adjustment operation information. If the difference between the adjustment time and the time of the conflict in the adjustment operation information is within a preset difference range, then the operation corresponding to the adjustment operation information is marked as a related path adjustment event. By adjusting events according to relevant paths, the extent of the diffusion of the cumulative effect of conflict can be obtained.

[0013] To address the same technical problem, a second aspect of this invention provides a supply chain anomaly intelligent identification system based on multi-dimensional data, comprising: The acquisition module is used to acquire order information and logistics information for each order in the order management system. The first analysis module is used to obtain the delay duration based on the timestamps corresponding to the order information and logistics information of each order, and to obtain the inventory allocation anomaly characteristics based on the delay duration. The inventory allocation anomaly characteristics include frequency distribution and high-frequency deviation range. The second analysis module is used to analyze cross-system interface fault logs based on frequency distribution and high-frequency deviation range to obtain a set of status field conflict points. The third analysis module is used to obtain the adjustment operation information of conflicting orders based on the conflict point set of the status field, analyze the adjustment operation information to obtain the degree of diffusion, and analyze the interaction logs based on the degree of diffusion to obtain the root cause distribution of order processing conflicts. The anomaly identification module is used to adjust and analyze the fusion parameters of each order according to the root cause distribution, obtain the delay impact path diagram, and obtain the anomaly cause identification result based on the delay impact path diagram.

[0014] The technical solution of this invention has the following advantages: The intelligent supply chain anomaly identification method based on multi-dimensional data provided in this invention obtains order information and logistics information for each order in an order management system; based on the timestamps corresponding to the order information and logistics information of each order, the delay duration is obtained; based on the delay duration, inventory allocation anomaly characteristics are obtained, including frequency distribution and high-frequency deviation range; the cross-system interface fault logs are analyzed based on the frequency distribution and high-frequency deviation range to obtain a set of status field conflict points; based on the set of status field conflict points, the adjustment operation information of conflicting orders is obtained; the adjustment operation information is analyzed to obtain the diffusion degree; based on the diffusion degree, interaction logs are analyzed to obtain the root cause distribution of order processing conflicts; the fusion parameters of each order are adjusted and analyzed based on the root cause distribution to obtain a delay impact path diagram; based on the delay impact path diagram, the anomaly cause identification result is obtained. This method improves the accuracy of supply chain anomaly identification through in-depth analysis of cross-system data. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart of the intelligent identification method for supply chain anomalies based on multi-dimensional data in an embodiment of the present invention; Figure 2 This is a system block diagram of the intelligent supply chain anomaly identification system based on multi-dimensional data in an embodiment of the present invention; Figure reference numerals: 200, Intelligent supply chain anomaly identification system based on multi-dimensional data; 201, acquisition module; 202, first analysis module; 203, second analysis module; 204, third analysis module; 205, anomaly identification module. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] In the description of this invention, it should be noted that the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] The intelligent supply chain anomaly identification method based on multi-dimensional data provided in this invention embodiment, such as... Figure 1 The diagram shows a flowchart of a supply chain anomaly intelligent identification method based on multi-dimensional data, including steps S101 to S104, each step of which is detailed below: S101: Obtain order information and logistics information for the order to be evaluated.

[0020] In this embodiment, the order creation timestamp, logistics status update timestamp, and current order status field are obtained from the order management database and the logistics tracking database. Order information refers to information obtained from the order management system, such as the time the order was generated in the system. Logistics information refers to information obtained from the logistics tracking platform, including the status update time, such as the time the order was marked as accepted in the logistics system; and the current order status field, such as "awaiting shipment" or "awaiting transportation".

[0021] S102: Based on the timestamps corresponding to the order information and logistics information, obtain the delay duration, and based on the delay duration, obtain the frequency distribution and deviation magnitude of inventory allocation anomalies.

[0022] In this embodiment, after obtaining order information and logistics information, the order creation time and logistics status update time are compared based on the timestamps of the order information and logistics information, namely the order creation time and the logistics status update time, to determine whether the data synchronization of the order to be evaluated is abnormal. By judging multiple orders to be evaluated and analyzing the time distribution pattern of data synchronization abnormality, the abnormal characteristics of inventory allocation are analyzed based on the obtained time distribution pattern.

[0023] In one embodiment, inventory allocation anomaly characteristics are obtained based on the delay duration, including: Determine if the delay duration is greater than a preset threshold. If the delay duration is greater than the preset threshold, mark the order as having a data synchronization error. Analyze the frequency of orders that experience data synchronization anomalies in different time periods to obtain the anomaly trigger frequency; Obtain inventory change logs within a time period where the frequency of abnormal triggering is greater than a preset frequency, extract and analyze the inventory change logs to obtain inventory allocation anomaly characteristics, wherein the inventory allocation anomaly characteristics include frequency distribution and high-frequency deviation range.

[0024] In this embodiment, the difference between the order creation time and the logistics status update time, i.e., the delay duration, is calculated. If the difference exceeds a preset duration threshold, such as 60 seconds, the order to be evaluated is marked as having a data synchronization anomaly. The percentage of anomaly records and the anomaly trigger frequency are statistically analyzed by time period (e.g., every 2 hours) to identify the time periods with concentrated anomalies. For example, monitoring the order management system reveals that from 9:00 AM to 11:00 AM, the anomaly frequency is 15%, with 15 out of 100 orders experiencing synchronization anomalies; from 2:00 PM to 4:00 PM, the anomaly frequency is 8%, with 8 out of 100 orders experiencing synchronization anomalies, indicating that the anomalies are concentrated between 9:00 AM and 11:00 AM. Then, the inventory change logs for the time periods with concentrated anomalies, such as 9:00 AM to 11:00 AM, are retrieved to extract key information, such as changes in inventory quantity (e.g., from 100 items to 95 items), allocation time (e.g., the time when inventory is allocated to an order), and the allocated order number. The preset frequency can be determined according to actual needs; it can be the value with the highest anomaly trigger frequency in each time period, or it can be determined based on actual requirements.

[0025] In one embodiment, the inventory change log is extracted and analyzed to obtain abnormal inventory allocation characteristics, including: Extract inventory change logs to obtain inventory change information, and calculate the correlation coefficient between inventory update lag time and the number of order allocation errors based on the inventory change information; Based on the correlation coefficient, abnormal records where the current inventory quantity does not match the inventory quantity occupied by the order are obtained, and the absolute value of the deviation for each abnormal record is calculated. The distribution density of the absolute value of the statistical deviation is used to obtain the frequency of anomalies in each deviation interval. The distribution ratio of the frequency of anomalies in different deviation intervals is calculated. Based on the distribution ratio, the range of high-frequency deviations in inventory allocation errors is determined, and the abnormal characteristics of inventory allocation are obtained.

[0026] In this embodiment, key information is extracted, such as changes in inventory quantity (e.g., from 100 to 95 units), allocation time (e.g., the time when inventory is allocated to an order), and the allocated order number. The preset frequency can be determined based on actual needs; it can be the value with the highest frequency of abnormal triggers in each time period, or determined according to actual requirements. Based on the key information in the inventory change log, the correlation coefficient between inventory update lag time and the number of order allocation errors is calculated. If the correlation coefficient exceeds a preset correlation threshold, it is determined that inventory operation delay has a significant impact on order fulfillment accuracy. It should be noted that after statistically analyzing 100 similar cases, the Pearson correlation coefficient between inventory update lag time and the number of order allocation errors reached 0.78. Therefore, 0.7 is set as the threshold for the correlation coefficient. When it is greater than 0.7, it indicates a strong correlation between inventory operation delay and order fulfillment accuracy. The calculation results of the correlation coefficient provide a scientific basis for the subsequent identification of abnormal records. After confirming the strong correlation, the system further filters specific cases of inconsistent inventory data.

[0027] Inventory update lag time can be understood as how long it takes for the inventory system to update when an order requires inventory. Order allocation errors can be understood as assigning 3 items when 5 should have been allocated. For example, if order synchronization is slow between 9-11 AM, we can also check if the inventory system is slow during this period. If the slower the inventory update and the more order allocation errors, it indicates that slow inventory updates are one of the main causes of order errors. Further analysis of inventory allocation deviation characteristics reveals abnormal inventory allocation characteristics, including frequency distribution and deviation magnitude.

[0028] Based on key information in the inventory logs, abnormal records where the current inventory quantity does not match the inventory quantity occupied by orders are extracted. The absolute value of the deviation for each abnormal record is calculated, and then the distribution ratio of the abnormal frequency in different deviation intervals is calculated. Based on the distribution ratio, the high-frequency deviation range of inventory allocation errors is determined, i.e., the abnormal frequency of each interval / the sum of the abnormal frequencies of all intervals, to obtain the proportion of each interval to the total abnormalities. For example, at a certain moment, the inventory management system shows that the current inventory of a certain product is 50 units, but the inventory occupied by pending orders in the order system has reached 65 units, a difference of 15 units. By collecting similar abnormal records and calculating the absolute value of the deviation, it was found that the deviation values ​​are mainly concentrated in the 5-20 unit range, which contains 70% of the total abnormal records, while large deviations exceeding 50 units account for only 5%. This distribution density characteristic reveals the main risk range of inventory allocation deviations. Therefore, 5-20 units is the high-frequency deviation range, and the corresponding frequency distribution is obtained.

[0029] The distribution analysis of anomaly frequencies further validated the regularity of the deviation magnitude. By statistically analyzing the number of anomaly records within different deviation ranges, it was found that the anomaly frequency was 3.2 times per hour for the 5-10 item deviation range, 1.8 times per hour for the 10-20 item deviation range, and decreased to 0.5 times per hour for the 20-50 item deviation range. This decreasing frequency distribution pattern indicates that small deviations are the main manifestation of inventory allocation errors, while large deviations are low-frequency but high-risk anomalies.

[0030] S103: Analyze the cross-system interface fault logs based on frequency distribution and deviation magnitude to obtain a set of conflict points in the status fields.

[0031] In this embodiment, the time point and impact range of the delivery trajectory deviation are determined based on the frequency distribution and high-frequency deviation range. Then, based on the time point and impact range of the delivery trajectory deviation, the order processing conflict and status update omission records are compared to obtain the conflict comparison distribution. Combined with the status update deviation coefficient calculated by analyzing the cross-system interface fault log, the set of conflict points where the status fields are inconsistent is determined, that is, the status field conflict point set.

[0032] In one embodiment, cross-system interface fault logs are analyzed based on frequency distribution and high-frequency deviation range to obtain a set of status field conflict points, including: Based on the frequency distribution and high-frequency deviation range, the delivery information of the order to be evaluated is obtained; By comparing and analyzing delivery information, the time points of delivery trajectory deviation and the geographical scope of impact can be obtained; By analyzing the time points of delivery trajectory deviations and the geographical impact range, a set of conflict points in status fields can be obtained from the cross-system interface fault logs.

[0033] In this embodiment, based on the abnormal characteristics of inventory allocation, including frequency distribution and high-frequency deviation range, the corresponding order's GPS coordinates, transport vehicle number, delivery time node, and other fields are obtained from the logistics trajectory database, i.e., delivery information. Then, the spatial distance is calculated by comparing the GPS coordinates with the warehouse location coordinates in the inventory record. If the spatial distance exceeds the preset delivery radius threshold, the corresponding order is marked as a location matching abnormality, and the spatial association verification result between the actual location information and the inventory record is obtained, i.e., the abnormal result.

[0034] Then, based on the spatial association verification results, the coordinates of the pre-determined route, the coordinates of the actual driving route, and the navigation route data of the delivery vehicle are extracted from the logistics tracking platform. The trajectory deviation value is calculated by the coordinate difference between the pre-determined route and the actual route. The time point and scope of the delivery trajectory deviation are determined. The conflict comparison distribution is obtained by comparing the order processing conflict and the missing status update records. The status update deviation coefficient is calculated by analyzing the cross-system interface fault logs to determine the set of conflict points where the status fields are inconsistent.

[0035] In one embodiment, comparative analysis of delivery information is performed to obtain the time points and impact range of delivery trajectory deviations, including: The spatial distance is calculated by comparing the transportation coordinates in the delivery information with the warehouse location of the order to be evaluated. The delivery information includes the transportation coordinates, the vehicle number, and the delivery time. If the spatial distance is greater than the preset delivery radius threshold, the location matching is determined to be abnormal, and an abnormal result is obtained. Based on the abnormal result, the vehicle's predetermined path coordinates, actual driving path coordinates, and navigation path data are extracted from the logistics tracking platform. Based on the coordinate difference between the predetermined path and the actual path, the trajectory deviation value is obtained. Based on the trajectory deviation value, the time point of the delivery trajectory deviation and the geographical impact range are obtained.

[0036] In this embodiment, spatial distance is calculated using GPS coordinates and warehouse location coordinates in the inventory record. If the spatial distance exceeds a preset delivery radius threshold, the corresponding order is marked as having an abnormal location match, thus obtaining the spatial association verification result between the actual location information and the inventory record, i.e., the abnormal result. For example, the GPS trajectory starting coordinates of an order are 31.2304°N, 121.4737°E, while the corresponding inventory record shows the warehouse location as 31.2315°N, 121.4742°E, with a spatial distance of approximately 150 meters between the two points. When the preset delivery radius threshold is 500 meters, this distance is within a reasonable range and is marked as a normal location match. However, when it is found that the distance between the GPS starting point of an order and the warehouse location in the inventory record reaches 2.3 kilometers, the system marks it as having an abnormal location match. This abnormality usually indicates that there is a problem of incorrect warehouse selection or delayed inventory record updates during the order allocation process.

[0037] Then, based on the spatial correlation verification results, the coordinates of the predetermined route, the coordinates of the actual driving route, and the navigation route data of the delivery vehicle are extracted from the logistics tracking platform. The trajectory deviation value is calculated by the coordinate difference between the predetermined route and the actual route. Based on the trajectory deviation value, the time point and geographical impact range of the delivery trajectory deviation are obtained.

[0038] In one embodiment, based on the trajectory deviation value, the time point of delivery trajectory deviation and the geographical impact range are obtained, including: Anomaly trajectory pattern analysis is performed based on trajectory deviation values ​​to obtain spatial distribution characteristics and deviation degree characteristics of trajectory deviation. Spatial distribution features are extracted to obtain multiple timestamps corresponding to each coordinate point; Calculate the difference between any two timestamps to obtain multiple misalignment magnitudes. If the misalignment magnitude is greater than a preset time deviation threshold, the timestamp misalignment is determined to be abnormal. Calculate the distance between the current vehicle position when the timestamp misalignment occurs and the vehicle position collected by the GPS device when the timestamp misalignment occurs, and obtain the distance difference. If the distance difference exceeds the preset location error range, it is judged as a location matching anomaly, and the time point of trajectory deviation and geographical impact range are determined.

[0039] In this embodiment, based on the trajectory deviation values ​​of each order, a clustering method is used to identify abnormal trajectory patterns in path deviation judgment, determining the spatial distribution and degree characteristics of trajectory deviation. For example, the optimal delivery route planned by the navigation system includes 15 key coordinate points, while the actual vehicle trajectory records 23 coordinate points. The additional 8 coordinate points reflect the driver's path deviation behavior. By calculating the lateral offset distance between corresponding coordinate points, it was found that the offset distance at the 7th coordinate point reached 800 meters, far exceeding the normal deviation range of 200 meters. By classifying similar deviation patterns using the path clustering method, three main abnormal trajectory types were identified: "detour to avoid congestion," "address error," and "equipment failure." This classification result provides basic data support for subsequent deviation cause analysis.

[0040] Based on the spatial distribution characteristics of trajectory deviation, the data recording timestamp, GPS device timestamp, and platform receiving timestamp corresponding to each coordinate point in the logistics tracking platform are extracted. Timestamp misalignment is identified by calculating the difference between these timestamps. If the timestamp misalignment exceeds a preset time deviation threshold, it is judged as a timestamp misalignment anomaly, resulting in a time distribution pattern for timestamp misalignment anomalies. In logistics tracking, timestamp misalignment manifests as a disconnect between the data recording time and the actual event occurrence time. For example, the GPS device collects vehicle location information at 10:25:33, but the logistics platform's data recording timestamp shows 10:27:18, a difference of 105 seconds. When there is a delay in the data transmission network, the platform's receiving timestamp may be further delayed to 10:28:45, forming a three-layer timestamp misalignment. Statistical analysis shows that timestamp misalignments within 60 seconds account for 85% of all records, while severe misalignments exceeding 300 seconds account for only 3%.

[0041] Then, based on the time distribution pattern of timestamp misalignment anomalies combined with actual location coordinates, the accuracy of the delivery vehicle's geographical location at the misaligned time point is verified. If the distance between the vehicle's location recorded by the misaligned timestamp and the vehicle's location collected by the GPS device at the same moment exceeds a preset location error range, it is judged as a location matching anomaly. The time point and geographical impact range of the trajectory deviation are determined, and the spatiotemporal distribution characteristics of the trajectory deviation event are obtained. The location verification process confirms the vehicle's true geographical location by cross-comparing multiple timestamp records. For example, when a timestamp misalignment record of 10:25:33 shows the vehicle's location as a commercial area, the system simultaneously retrieves the original data collected by the GPS device at that moment and finds that the vehicle's actual location is in a residential area 1.2 kilometers away from the commercial area, indicating a mismatch in location information. This directly affects the customer's accurate understanding of the delivery progress and also increases the difficulty of investigating delivery anomalies.

[0042] In one embodiment, cross-system interface fault logs are analyzed based on the delivery trajectory deviation time points and geographical impact range to obtain a set of status field conflict points, including: Obtain order change status information, filter the order change status information based on the trajectory deviation time point, and obtain the order processing event; If the difference between the timestamp of the order processing event and the time point of the delivery trajectory deviation falls within a preset time window, the order processing event will be marked as a conflict event. Based on the processing information and status update information of conflict events, the frequency of changes and the number of omissions of conflict events within the conflict period are obtained; Based on the frequency of changes and the number of omissions, a set of conflict points in the status fields is obtained.

[0043] In this embodiment, based on the delivery trajectory deviation time point and geographical impact range, order status change records, processing operation logs, and status field update timestamps for the corresponding time period are retrieved from the order processing database. All order processing events during the trajectory deviation period are extracted by filtering through the time range. If the difference between the timestamp of an order processing event and the trajectory deviation time point is within a preset time window, it is marked as a conflict event, thus obtaining conflict event records and status update omission event records. Based on these conflict event records and status update omission event records, database queries are used to statistically analyze the change frequency and omission count of different order status fields within the conflict time period. Furthermore, based on the change frequency and omission count within the conflict time period, a set of status field conflict points is obtained.

[0044] In one embodiment, a set of conflict points for the status field is obtained based on the frequency of changes and the number of omissions, including: Based on the frequency of changes and the number of omissions, the distribution characteristics of conflict events in the time and business dimensions are calculated to obtain the type classification and quantity distribution of conflict events; Based on the type classification and quantity distribution of conflict events, determine the conflict comparison distribution results; Based on the conflict comparison distribution results, interface data transmission information is extracted from the cross-interface communication logs, and the interface data transmission information is analyzed to obtain a set of conflict points in the status field.

[0045] In this embodiment, after obtaining the frequency of changes and the number of omissions within the conflict period, frequency statistics methods are used to calculate the distribution characteristics of various conflict events in the time and business dimensions, obtain the type classification and quantity distribution of conflict events, and determine the conflict comparison distribution results. For example, the identification of trajectory deviation time points establishes a time benchmark for screening order processing conflict events. When a delivery vehicle deviates from its predetermined route by more than 500 meters at 14:25:33, this time point is marked as an anomaly occurrence node. Through a time window mechanism, the system automatically extracts all relevant order processing events within the 10-minute time period from 14:20:00 to 14:30:00. Within this time window, three order status change records and two processing operation log anomalies were found. Among them, the status update request of a certain order at 14:26:15 was marked as a processing conflict event due to inaccurate location information caused by trajectory deviation. This time-related event screening mechanism can accurately locate business processing problems related to trajectory anomalies, providing an accurate data range for subsequent analysis. Based on the establishment of processing conflict event records, frequency statistics analysis reveals the distribution pattern of conflict events through quantitative methods.

[0046] Specifically, database queries revealed that in the past 24 hours, the "Delivery Status" field had 15 conflict occurrences, the "Inventory Status" field had 8, while the "Customer Confirmation" field had only 3. This distribution indicates that the delivery status field is most susceptible to trajectory deviations, making it a high-risk area for conflicts. Furthermore, statistics on missed events show that status update omissions were most concentrated between 2-4 PM, accounting for 40% of all missed events, a period that coincides with peak delivery times. By establishing a conflict comparison distribution model, managers can identify system weaknesses and high-risk periods.

[0047] It should be noted that the conflict comparison distribution results provide crucial business background information for interface fault analysis. The cross-interface communication log analysis process verifies the root causes of problems discovered at the business level from a technical perspective.

[0048] Based on the conflict comparison distribution results, interface call failure records, data transmission anomaly records, and interface response timeout records are extracted from the cross-interface communication logs. Correlation analysis is used to calculate the correlation coefficient between the frequency of interface failures and the number of status update failures. If the correlation coefficient exceeds a preset correlation threshold, the impact of interface failures on status updates is determined, resulting in a status update deviation coefficient. Based on the status update deviation coefficient and the actual and expected values ​​of the order status field, a field-by-field comparison is performed. Field value difference detection identifies the specific locations of inconsistencies in order status, inventory status, and delivery status. If the field difference exceeds a preset deviation range, it is identified as a status field conflict point. The locations of all status field inconsistencies are determined, resulting in a set of status field conflict points.

[0049] For example, when an abnormal frequency of conflicts in the delivery status field is detected, the system simultaneously checks the communication records between the order management interface and the logistics tracking interface. It finds that at 14:25:33, when the trajectory deviation occurred, the interface call timed out three times consecutively, each timeout lasting 8 seconds. Through correlation analysis, the correlation coefficient between the frequency of interface failures and the number of status update failures reached 0.82, exceeding the preset threshold of 0.75, confirming that the interface communication problem was the main technical cause of the status field conflicts. This dual technical-business verification mechanism improves the accuracy and reliability of problem diagnosis.

[0050] Furthermore, the application of the status update deviation coefficient guides the precise conflict localization process at the field level. When the deviation coefficient indicates that the impact of interface failure on status updates is at a medium risk level, the system initiates a field-by-field comparison and verification process. A detailed inspection of a particular order revealed that its "Current Location" field displayed "Transfer Station B," but the actual GPS trajectory data showed the vehicle was located at "Delivery Point C," a significant discrepancy. Simultaneously, the "Estimated Arrival Time" field displayed 15:30:00, but the recalculated estimated arrival time based on the current location and traffic conditions was 16:15:00, a 45-minute time deviation. By establishing a field difference detection mechanism, the system can accurately identify the location of each conflicting status field, forming a complete set of conflict points. This provides detailed data support for the precise repair of order statuses and system optimization, significantly improving the accuracy of order processing and customer service quality.

[0051] S104: Based on the set of conflict points in the status field, obtain the adjustment operation information of conflicting orders, analyze the adjustment operation information to obtain the degree of diffusion, and analyze the interaction logs based on the degree of diffusion to obtain the root cause distribution of order processing conflicts.

[0052] In this embodiment, temporary adjustment records are extracted from the logistics route planning system. Combined with a set of conflict points, the correlation between logistics route misjudgments and data overlay loss is analyzed. The correspondence between route changes before and after adjustments and timestamp misalignment records is derived, and the degree of diffusion of the cumulative conflict effect is determined. Based on the diffusion degree, the interaction logs of cross-system interface failures and system response delays are analyzed to obtain the failure interaction distribution. Abnormal records caused by heterogeneous data formats are extracted, and a correlation mapping between data synchronization delays and missing status updates is constructed to determine the root cause distribution of order processing conflicts.

[0053] Based on the root cause distribution, the fusion parameters of order status, inventory data, and logistics trajectory are adjusted. The time synchronization factor and data source priority of cross-system interfaces are dynamically updated. Combined with inventory update lag and delivery trajectory deviation triggering conditions extracted from historical records, the impact path of system response delay on anomaly triggering frequency is analyzed, generating a delay impact path diagram. Specifically, based on the diffusion degree of the conflict accumulation effect in the logistics network, interface call failure records, response timeout records, and data transmission error records are obtained from cross-interface communication logs. Simultaneously, interaction information such as interface response time, call frequency, and failure type is extracted. Time correlation matching is used to compare and verify the occurrence time of diffusion events with the occurrence time of interface failures. If the time difference is within a preset correlation window, it is marked as a related failure event, obtaining the interaction data between interface failures and response delays. Based on the interaction data between interface failures and response delays, the distribution characteristics of different failure types in time and space are calculated through failure type classification statistics. Frequency statistics methods are used to identify the interface locations and time periods where failures are concentrated, obtaining the distribution area and occurrence pattern of failure events, and determining the failure interaction distribution pattern. Based on the fault interaction distribution pattern, records of data format mismatch, field mapping errors, and encoding conversion failures are extracted from the data transmission logs. Data format comparison identifies transmission anomalies and parsing errors caused by data heterogeneity. If the time overlap between data format anomaly events and synchronization delay events exceeds a preset threshold, a correlation is established, resulting in a mapping between data synchronization delay and missing status updates. Based on this mapping, combined with historical records of order processing conflicts and fault occurrence patterns, cluster analysis is used to group similar conflict events according to their causes. The distribution ratio and impact of each root cause in the total conflict events are statistically analyzed to determine the distribution of interface faults, data heterogeneity, and synchronization delays as the main root causes, thus obtaining the root cause distribution of order processing conflicts. For example, when the cumulative effect of the conflict is rated as moderate, the cross-interface communication logs show 12 interface call failures within the same time period, with the order management interface and inventory management interface experiencing the most failures (7 times). Response timeout records show an average timeout of 15.6 seconds, far exceeding the normal response threshold of 5 seconds. Data transmission errors mainly fall into two categories: JSON parsing failures and XML data structure mismatches. Time correlation matching revealed that the overlap between the propagation event time of 14:25:33 and the concentrated occurrence of interface failures from 14:24:18 to 14:27:45 reached 89%, far exceeding the preset correlation threshold of 60%. This confirms that the interface failure was the direct technical cause of the conflict propagation.

[0054] In one embodiment, the adjustment operation information is analyzed to obtain the degree of diffusion, including: If there is a conflict, the order number in the conflict point set of the status field is matched with the order number in the adjustment operation information. If the difference between the adjustment time and the time of the conflict in the adjustment operation information is within a preset difference range, then the operation corresponding to the adjustment operation information is marked as a related path adjustment event. By adjusting events according to relevant paths, the extent of the diffusion of the cumulative effect of conflict can be obtained.

[0055] In this embodiment, the temporary route adjustment record, adjustment operation timestamp, and adjustment reason identifier for the corresponding order with a conflicting status field are retrieved from the logistics route planning database based on the conflict point set. The order numbers of each conflict point in the conflict point set are matched with the order numbers in the temporary route adjustment records. If the order numbers match and the difference between the adjustment time and the conflict occurrence time is within a preset time range, it is marked as a related route adjustment event. After judging all conflict points, the conflict-related route adjustment event record is obtained. For example, the matching association between the conflict point set and the route adjustment record establishes a precise data association foundation through the order number. When the order status field conflict point set contains the order number ORD20240315001, the corresponding temporary adjustment record in the route planning database shows that the order underwent an emergency route change at 14:25:33, with the adjustment reason identifier being "road congestion." Time matching verification reveals that the difference between the conflict occurrence time 14:23:18 and the route adjustment time is 2 minutes and 15 seconds, which is within the preset time range of 5 minutes; therefore, it is marked as a related route adjustment event. This dual association mechanism based on order number and time window ensures the accuracy of data matching, avoids interference from irrelevant adjustment events on the analysis results, and establishes a reliable data foundation for subsequent in-depth association analysis.

[0056] Based on the conflict-related route adjustment event records, records of route misjudgment operations, data overlay failures, and original route data loss are extracted. Database correlation queries identify the specific operation types that caused route data anomalies during the adjustment events. Frequency statistics are used to calculate the co-occurrence frequency of route misjudgment events and data overlay loss events to determine the strength of the correlation between logistics route misjudgment and data overlay loss. For example, the original route contained 15 key coordinate points, with the starting point at 31.23°N, 121.47°E and the ending point at 31.19°N, 121.51°E. The adjusted route contained 23 coordinate points, with the 8 newly added coordinate points mainly concentrated in the middle detour area, with a maximum positional offset of 1.8 kilometers. Timestamp misalignment records show that the route adjustment operation was recorded at 14:25:33, but the actual GPS track change time was 14:27:18, a time difference of 1 minute and 45 seconds. By establishing a time-space correspondence table, managers can accurately track the specific time and geographical location of each path change, providing crucial reference for the precise location of abnormal events.

[0057] Based on the correlation between logistics route misjudgment and data overlay loss, the coordinate sequences of the route before and after adjustment, as well as the timestamp records of route changes, are extracted. The positional offset and magnitude of the route change are calculated through coordinate point comparison. A matching relationship is established between the time of route change and the time of occurrence of the time anomaly, obtaining the correspondence data between route change and timestamp misalignment. Based on the correspondence data between route change and timestamp misalignment, the number of subsequent abnormal events and the number of affected orders triggered by a single conflict point are statistically analyzed. The propagation chain of abnormal events identifies the diffusion path and affected nodes of the conflict in related orders. If the number of subsequent abnormal events exceeds a preset diffusion threshold, it is judged as a high-impact conflict event, determining the degree of diffusion of the cumulative effect of the conflict.

[0058] The mechanism for identifying the propagation chain of abnormal events quantifies the scope of impact by tracing the diffusion path of conflicts. When order ORD20240315001 experienced a route adjustment conflict, its three associated subsequent orders also experienced delivery delays, creating a significant propagation effect. Statistical analysis of abnormal events revealed that a single route conflict triggered an average of 2.3 subsequent abnormal events, affecting an average of 4.7 orders. When the number of subsequent abnormal events for a conflict exceeds five, it is classified as a high-impact conflict event, requiring the initiation of emergency response procedures. This quantitative assessment mechanism of the degree of diffusion enables logistics managers to quickly identify high-risk conflict events, prioritize resource allocation for handling, effectively control the further spread of the cumulative effect of conflicts, and significantly improve the stability and resilience of the overall logistics network.

[0059] Based on the changes in logistics routes before and after the adjustment, the nodes, routes, and timing characteristics of the route changes are identified, and the timing, frequency, and associated conflict points of the misalignment are analyzed. The set of conflict points is integrated, and the interaction and scope of influence of each conflict point in the route adjustment are evaluated. The diffusion trend of the cumulative conflict effect on the logistics network is obtained, and the impact of the diffusion trend on adjacent nodes, route delays, and overall scheduling efficiency is analyzed. The degree of diffusion of the cumulative conflict effect is comprehensively determined.

[0060] Based on the diffusion degree of the cumulative effect of conflict, the coordinates of path nodes before and after adjustment, as well as the route connection relationships, are obtained from the logistics route database. Specific node locations and newly added / deleted route segments are identified through coordinate point comparison. A time series sorting method is used to extract the operation time series of the route adjustment and the order of node changes, resulting in path node change characteristics and route adjustment time series data. Based on the path node change characteristics and route adjustment time series data, the specific time point of the anomaly, the duration of the anomaly, and the number of orders affected by the anomaly are extracted from the timestamp misalignment records. The correlation between the misalignment occurrence time and the route adjustment time is calculated through time point matching. The frequency of misalignment events and the corresponding number of conflict points in different time periods are statistically analyzed to determine the time misalignment distribution pattern and conflict point correlation data. Based on the time misalignment distribution pattern and conflict point correlation data, all relevant conflict points are integrated to form a unified conflict point set. The mutual influence relationship between conflict points is established through correlation matrix calculation, quantifying the influence intensity and scope of each conflict point in the route adjustment process. If the influence intensity between conflict points exceeds a preset correlation threshold, it is judged as a strong correlation conflict relationship, resulting in a conflict point correlation network and interaction strength data. Based on the data on the conflict point association network and interaction strength, the propagation path and diffusion nodes of the conflict in the logistics network are tracked. Quantitative indicators of the impact of the diffusion on the delivery time delay of adjacent nodes, the reduction of route traffic efficiency, and the overall scheduling efficiency are calculated. The overall degree of diffusion impact is comprehensively calculated by the weighted summation method to determine the degree of diffusion of the cumulative effect of the conflict in the logistics network.

[0061] S105. Adjust and analyze the fusion parameters of order status, inventory data and logistics trajectory for each order according to the root cause distribution to obtain the delay impact path diagram. Based on the delay impact path diagram, obtain the abnormal cause identification result.

[0062] In this embodiment, based on the root cause distribution of order processing conflicts, the data configuration management module obtains the fusion parameter configurations for each order, such as the order status field weight, inventory data update frequency, and logistics trajectory collection interval. The fusion weight ratio and update priority of different data sources are adjusted by matching root cause types. If the proportion of interface failure root causes exceeds a preset threshold, the fusion weight of order status data is increased. If the proportion of data heterogeneity root causes is high, the verification frequency of inventory data is increased, thus obtaining the fusion parameter adjustment scheme.

[0063] The time synchronization factor and data source access priority sequence in the cross-interface communication configuration are updated according to the fusion parameter adjustment scheme. The time synchronization factor adjusts the data transmission interval between different interfaces, and a queue sorting method is used to reorder the access order of each data source. If the response time of a high-priority data source exceeds a preset latency threshold, it automatically switches to the next lower priority data source, thus obtaining the interface time synchronization configuration and data source priority configuration. For example, the time synchronization factor is adjusted from a uniform 1.0 to a differentiated setting: 0.8 for the order management interface, 1.2 for the inventory management interface, and 1.0 for the logistics tracking interface. This differentiated time synchronization factor means that the data transmission interval of the order management interface is shortened by 20%, while the transmission interval of the inventory management interface is extended by 20%. The queue sorting method reorders the data source access priority as follows: order status data, inventory change data, logistics location data, and historical data. When the response time of the highest priority order status data source exceeds a preset threshold of 8 seconds, it automatically switches to a backup data source to ensure the continuity and reliability of data acquisition.

[0064] The establishment of interface time synchronization configuration provides a time benchmark for the accurate identification of abnormal triggering conditions. The analysis of historical operation logs uses time series mining to discover the triggering patterns of abnormal events. Therefore, based on the interface time synchronization configuration and data source priority configuration, time records of inventory update lag events, trigger times of delivery trajectory deviation events, and changes in data synchronization intervals are extracted from historical operation logs. Time series matching is used to identify the correlation patterns between each parameter in the triggering conditions and the occurrence of abnormal events, and the influence coefficient of interface response latency on the frequency of abnormal triggering is calculated to determine the relationship between abnormal triggering conditions and response latency. For example, statistics on inventory update lag events show that when the data synchronization interval exceeds 8 minutes, the probability of lag events increases sharply from the normal 2% to 15%. Analysis of the trigger time of delivery trajectory deviation events shows that when the interface response latency exceeds 12 seconds, the triggering frequency of trajectory deviation is 3.2 times higher than normal. Time series matching reveals that the setting of the abnormal triggering threshold directly affects the sensitivity of abnormal detection: a threshold that is too low will produce too many false alarms, while a threshold that is too high may miss real abnormal events. The calculation results of the impact coefficient show that for every 1 second increase in interface response latency, the average frequency of exception triggering increases by 0.15 times per hour. Based on the relationship between anomaly triggering conditions and response delay, a shortest path algorithm is used to construct the associated nodes and impact edges between response delay and anomaly triggering frequency. The propagation path of delay impact across different business modules is tracked through node connections. The impact strength and propagation time on each path are calculated. If the impact strength exceeds a preset propagation threshold, it is marked as a critical impact path, resulting in a delay impact path graph. The shortest path algorithm is a classic algorithm in graph theory used to find the shortest connection path between two nodes in a graph. In the delay impact analysis, each business module is abstracted as a node in the graph, and the data dependencies between modules are represented as directed edges. When a 5-second response delay occurs in the order management module, the impact first propagates to the inventory management module, increasing the delay to 7 seconds, then to the logistics tracking module, accumulating a delay to 10 seconds, and finally affecting the customer notification module, with a total delay of 13 seconds. Tracking the node connections reveals three main impact propagation paths: the order-inventory-logistics path has an impact strength of 0.85, the order-payment-finance path has an impact strength of 0.72, and the order-customer service-notification path has an impact strength of 0.63. When the impact intensity exceeds a preset propagation threshold of 0.8, the path is marked as a critical impact path, requiring focused monitoring and optimization. The construction of a delay impact path diagram allows managers to intuitively understand the propagation process of delays throughout the entire business system, providing a scientific basis for targeted performance optimization.

[0065] The intelligent supply chain anomaly identification system based on multi-dimensional data provided in this invention embodiment, such as... Figure 2 As shown, Figure 2 The system block diagram of the supply chain anomaly intelligent identification system 200 based on multi-dimensional data includes: The acquisition module 201 is used to acquire order information and logistics information of each order in the order management system; The first analysis module 202 is used to obtain the delay duration based on the timestamps corresponding to the order information and logistics information of each order, and to obtain the inventory allocation anomaly characteristics based on the delay duration. The inventory allocation anomaly characteristics include frequency distribution and high-frequency deviation range. The second analysis module 203 is used to analyze the cross-system interface fault logs based on the frequency distribution and high-frequency deviation range to obtain a set of status field conflict points. The third analysis module 204 is used to obtain the adjustment operation information of conflicting orders based on the conflict point set of the status field, analyze the adjustment operation information to obtain the degree of diffusion, and perform interaction log analysis based on the degree of diffusion to obtain the root cause distribution of order processing conflicts. The anomaly identification module 205 is used to adjust and analyze the fusion parameters of each order according to the root cause distribution, obtain the delay impact path diagram, and obtain the anomaly cause identification result based on the delay impact path diagram.

[0066] The specific implementation of the intelligent identification system for supply chain anomalies based on multi-dimensional data is basically the same as the specific implementation of the method for intelligent identification system for supply chain anomalies based on multi-dimensional data described above, and will not be repeated here.

[0067] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0068] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for intelligent identification of supply chain anomalies based on multi-dimensional data, characterized in that, include: Retrieve order information and logistics information for each order in the order management system; Based on the timestamps corresponding to the order information and logistics information of each order, the delay duration is obtained, and it is determined whether the delay duration is greater than a preset duration threshold. If the delay duration is greater than the preset duration threshold, the order is marked as a data synchronization anomaly. Analyze the frequency of orders that experience data synchronization anomalies in different time periods to obtain the anomaly trigger frequency; Obtain inventory change logs within a time period where the abnormal trigger frequency is greater than a preset frequency, extract and analyze the inventory change logs to obtain inventory allocation anomaly features, wherein the inventory allocation anomaly features include frequency distribution and high-frequency deviation range; Based on the frequency distribution and the high-frequency deviation range, the cross-system interface fault logs are analyzed to obtain a set of status field conflict points; Based on the set of conflict points in the status field, obtain the adjustment operation information of the conflicting orders, analyze the adjustment operation information to obtain the degree of diffusion, and perform interaction log analysis based on the degree of diffusion to obtain the root cause distribution of order processing conflicts. The fusion parameters of each order are adjusted and analyzed according to the root cause distribution to obtain a delay impact path diagram. Based on the delay impact path diagram, the abnormal cause identification result is obtained.

2. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 1, characterized in that, The extraction and analysis of inventory change logs yields abnormal inventory allocation characteristics, including: The inventory change log is extracted to obtain inventory change information. Based on the inventory change information, the correlation coefficient between inventory update lag time and the number of order allocation errors is calculated. Based on the correlation coefficient, abnormal records where the current inventory quantity is inconsistent with the inventory quantity occupied by the order are obtained, and the absolute value of the deviation for each abnormal record is calculated. The distribution density of the absolute value of the deviation is statistically analyzed to obtain the frequency of abnormalities in each deviation interval. The distribution ratio of the frequency of abnormalities in different deviation intervals is calculated. Based on the distribution ratio, the high-frequency deviation range of inventory allocation errors is determined, and the abnormal characteristics of inventory allocation are obtained.

3. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 1, characterized in that, The analysis of cross-system interface fault logs based on the frequency distribution and the high-frequency deviation range yields a set of status field conflict points, including: Based on the frequency distribution and the high-frequency deviation range, the delivery information of the order to be evaluated is obtained; By comparing and analyzing the delivery information, the time points of delivery trajectory deviation and the geographical impact range can be obtained; Based on the time points of the delivery trajectory deviation and the geographical impact range, the cross-system interface fault logs are analyzed to obtain a set of status field conflict points.

4. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 3, characterized in that, The comparative analysis of the delivery information yields the time points and geographical impact ranges of delivery trajectory deviations, including: The spatial distance is obtained by calculating the distance between the transportation coordinates in the delivery information and the warehouse location of the order to be evaluated. The delivery information includes the transportation coordinates, the vehicle number, and the delivery time. If the spatial distance is greater than the preset delivery radius threshold, the location matching is determined to be abnormal, and an abnormal result is obtained. Based on the abnormal result, the vehicle's predetermined path coordinates, actual driving path coordinates, and navigation path data are extracted from the logistics tracking platform. Based on the coordinate difference between the predetermined path and the actual path, a trajectory deviation value is obtained. Based on the trajectory deviation value, the time point and geographical impact range of the delivery trajectory deviation are obtained.

5. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 4, characterized in that, The process of obtaining the delivery trajectory deviation time point and geographical impact range based on the trajectory deviation value includes: Based on the trajectory deviation values, abnormal trajectory pattern analysis is performed to obtain the spatial distribution characteristics and deviation degree characteristics of the trajectory deviation. The spatial distribution features are extracted to obtain multiple timestamps corresponding to each coordinate point; Calculate the difference between any two timestamps to obtain multiple misalignment amplitudes. If the misalignment amplitude is greater than a preset time deviation threshold, the timestamp misalignment is determined to be abnormal. Calculate the distance between the current vehicle position at the time of the timestamp misalignment and the vehicle position collected by the GPS device at the time of the timestamp misalignment, and obtain the distance difference; If the distance difference exceeds the preset position error range, it is determined to be an abnormal position matching, and the time point of trajectory deviation and geographical impact range are determined.

6. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 3, characterized in that, The method involves analyzing cross-system interface fault logs based on the delivery trajectory deviation time points and the geographical impact range to obtain a set of status field conflict points, including: Obtain order change status information, and filter the order change status information according to the trajectory deviation time point to obtain multiple order processing events; Based on the timestamp of the order processing event and the time point of the delivery trajectory deviation, each order processing event is judged to obtain a record of processing conflict events and a record of missing status update events; Based on the conflict event records and the status update omission event records, the change frequency and omission count of each order during the conflict period are obtained. Based on the frequency of changes and the number of omissions, a set of conflict points in the status field is obtained.

7. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 6, characterized in that, The set of status field conflict points is obtained based on the change frequency and the number of omissions, including: Based on the frequency of changes and the number of omissions, the distribution characteristics of the conflict events in the time and business dimensions are calculated to obtain the type classification and quantity distribution of the conflict events; Based on the type classification and quantity distribution of the conflict events, the conflict comparison distribution results are determined; Based on the conflict comparison distribution results, interface data transmission information is extracted from the cross-interface communication logs, and the interface data transmission information is analyzed to obtain a set of status field conflict points.

8. The intelligent supply chain anomaly identification method based on multi-dimensional data as described in claim 1 or 7, characterized in that, The analysis of the adjustment operation information to obtain the diffusion degree includes: If there is a conflict, the order number in the conflict point set of the status field is matched with the order number in the adjustment operation information. If the difference between the adjustment time and the conflict occurrence time in the adjustment operation information is within a preset difference range, then the operation corresponding to the adjustment operation information is marked as a related path adjustment event. The extent of the cumulative effect of the conflict is obtained by adjusting the events according to the relevant paths.

9. A supply chain anomaly intelligent identification system based on multi-dimensional data, characterized in that, include: The acquisition module is used to acquire order information and logistics information for each order in the order management system. The first analysis module is used to obtain the delay duration based on the timestamps corresponding to the order information and logistics information of each order, determine whether the delay duration is greater than a preset duration threshold, and mark the order as a data synchronization anomaly if the delay duration is greater than the preset duration threshold. Analyze the frequency of orders that experience data synchronization anomalies in different time periods to obtain the anomaly trigger frequency; Obtain inventory change logs within a time period where the abnormal trigger frequency is greater than a preset frequency, extract and analyze the inventory change logs to obtain inventory allocation anomaly features, wherein the inventory allocation anomaly features include frequency distribution and high-frequency deviation range; The second analysis module is used to analyze the cross-system interface fault logs based on the frequency distribution and the high-frequency deviation range to obtain a set of status field conflict points. The third analysis module is used to obtain the adjustment operation information of the conflicting orders based on the conflict point set of the status field, analyze the adjustment operation information to obtain the diffusion degree, and perform interaction log analysis based on the diffusion degree to obtain the root cause distribution of order processing conflicts. An anomaly identification module is used to adjust and analyze the fusion parameters of each order according to the root cause distribution to obtain a delay impact path diagram, and based on the delay impact path diagram, obtain the anomaly cause identification result.

Images