A logistics management method and system based on a data platform

Abstract

The application relates to the technical field of logistics management, and discloses a logistics management method and system based on a data middle station, which comprises the following steps: acquiring and standardizing order information, warehouse state, vehicle position, customer feedback and settlement data; establishing a hierarchical structure in the data middle station and extracting indexes such as transportation timeliness, warehouse turnover, delivery track deviation and customer feedback grade; comparing the indexes with historical data to generate abnormality detection results; when there is an abnormality, introducing multi-source information for consistency verification and generating a candidate path; sequentially screening the final scheduling path through a time threshold, resource carrying capacity and track consistency; combining customer feedback to generate a compensation factor to correct operation indexes during execution; when the abnormality still exists, triggering path reconstruction and resource reallocation; and finally generating and displaying graded early warning information. Through comparison and path screening, the application improves the abnormality identification accuracy and the reliability of logistics scheduling.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, and more specifically, to a logistics management method and system based on a data platform. Background Technology

[0002] With the rapid development of e-commerce and supply chain management, modern logistics systems are becoming increasingly complex, involving multiple stages such as orders, warehousing, transportation, delivery, and customer service, generating massive amounts of multi-source, heterogeneous business data. Traditional logistics management systems typically employ independent and decentralized architectures for data processing; for example, order management systems, warehouse management systems, and transportation management systems operate independently, leading to inconsistent data standards and severe information silos. This data fragmentation makes it difficult for enterprises to achieve unified and real-time perception and insight into the operational status of the entire logistics chain. When detecting anomalies, existing methods often rely on indicators from a single system or a single dimension, lacking cross-validation from multi-source information. This makes them prone to misjudgments due to local fluctuations or occasional factors, triggering unnecessary scheduling instructions and increasing operating costs.

[0003] When anomalies are detected and route scheduling optimization is required, traditional scheduling scheme generation often relies on historical experience or static rules, resulting in a delayed response and an inability to fully integrate dynamic external information such as real-time road conditions, weather, and vehicle sensor readings. This leads to recommended scheduling routes having low feasibility and poor adaptability in actual execution. Furthermore, existing systems generally lack mechanisms to effectively integrate end-customer feedback into the operational indicator correction and scheduling decision-making closed loop, causing a disconnect between logistics service quality assessment and operational optimization, making it difficult to achieve refined and intelligent management centered on customer experience. Summary of the Invention

[0004] In view of this, the present invention proposes a logistics management method and system based on a data middle platform, aiming to solve the problems of high error rate and lagging scheduling scheme caused by the difficulty of data dispersion and integration in the prior art.

[0005] In one aspect, this invention proposes a logistics management method based on a data platform, comprising: Acquire business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and standardize the business data to obtain standard business data; The standard business data is managed hierarchically in the data platform. The hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the hierarchical process, cleaning, noise reduction, and format unification are completed to obtain hierarchical data. Logistics operation indicators are extracted from the hierarchical data. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. The logistics operation indicators are then compared with a set of historical indicators to obtain anomaly detection results. When the anomaly detection result indicates an anomaly, a consistency verification is performed based on vehicle sensor information, weather information, and traffic information. When the consistency verification result is anomaly, a set of candidate scheduling paths is generated. The candidate scheduling path set is subjected to hierarchical screening, which is carried out by time threshold, resource carrying capacity and path trajectory consistency in turn, to obtain the final scheduling path; Logistics scheduling is implemented according to the final scheduling path, and a compensation factor is generated based on the customer feedback level to correct the logistics operation indicators during the execution process. When the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation are triggered in sequence. Based on the revised logistics operation indicators, a tiered early warning information is generated, which is divided into Level 1, Level 2 and Level 3 early warnings, and triggers corresponding risk warnings and handling procedures. The logistics operation indicators, candidate scheduling path set, corrected logistics operation indicators, and hierarchical early warning information are displayed.

[0006] Furthermore, when standardizing the business data to obtain standard business data, the process includes: The time field of the business data is formatted uniformly, and time records from different sources are converted into a unified timestamp format. The numerical fields in the aforementioned business data shall be standardized to the International System of Units (SI). The text fields of the business data are unified to UTF-8 encoding to obtain standard business data.

[0007] Furthermore, when managing the standard business data in a hierarchical manner within the data platform, the following steps are included: When establishing the operational data layer, the standard business data is stored one by one according to the field dimension, and missing value supplementation, abnormal record removal and duplicate record deduplication are performed during the storage process to obtain the operational layer data; When establishing the detailed data layer, the operation layer data is integrated according to business objects and time series. During the integration process, noise reduction and format unification are performed on cross-system records of the same business object to generate detailed data. In the process of establishing the data mart layer, the detailed data is used as the source and aggregated according to transportation timeliness, warehouse turnover, delivery trajectory and customer feedback. During the aggregation process, the standardization of indicators from different sources is adjusted to form hierarchical data.

[0008] Furthermore, when extracting logistics operation indicators from the layered data and comparing them with a historical indicator set, the process includes: The transportation timeliness is compared with the historical average. When the difference exceeds the time threshold, the transportation timeliness comparison result is marked as abnormal. The time threshold is obtained based on the statistical distribution range of historical transportation timeliness data. The warehouse turnover rate is compared with the historical average. When the difference exceeds the turnover threshold, the warehouse turnover comparison result is marked as abnormal. The turnover threshold is obtained based on the fluctuation range of the historical warehouse turnover rate. The delivery trajectory deviation is compared with the historical trajectory range. When the delivery trajectory deviation is not within the historical trajectory range, the delivery trajectory comparison result is marked as abnormal. The historical trajectory range is set by the spatial coverage range of multi-day delivery trajectory data. The customer feedback level is compared with the historical distribution range. When the customer feedback level is low and below the lower limit of the historical distribution range, the customer feedback comparison result is marked as abnormal. The historical distribution range is divided into upper limit range, middle range and lower limit range based on the statistical results of customer feedback level in several historical periods. An anomaly detection result is generated when any of the comparison results of transportation timeliness, warehousing turnover, delivery trajectory, and customer feedback is marked as abnormal.

[0009] Furthermore, when the anomaly detection result indicates the presence of an anomaly, a consistency verification is performed, including: Based on weather information, real-time temperature, humidity, and precipitation type are extracted, and the presence of meteorological conditions that could cause transportation delays is determined based on temperature fluctuations, humidity trends, and precipitation conditions, thus obtaining the verification results of the weather information. When both the verification results of the weather information and the abnormal comparison results of transportation timeliness indicate that there is a transportation delay, the transportation timeliness comparison results are marked as consistent anomalies. Based on traffic information, the road congestion index, road segment travel time and historical average are extracted, and the presence of traffic conditions that lead to an extension of the cargo turnover cycle is determined based on road congestion and travel time, thus obtaining the verification results of traffic information; when both the verification results of traffic information and the abnormal comparison results of warehouse turnover determine that the cargo turnover cycle is extended, the warehouse turnover comparison results are marked as consistent anomalies. Based on vehicle sensor information, the vehicle's speed, acceleration, and positioning trajectory are extracted. The presence of path deviation is determined based on speed fluctuations, acceleration changes, and positioning trajectory offsets, thus obtaining the verification results of the vehicle sensor information. When both the verification results of the vehicle sensor information and the abnormal comparison results of the delivery trajectory deviation indicate that there is a path deviation, the delivery trajectory deviation comparison results are marked as consistent anomalies. When the verification results of two or more types of information, including weather information, traffic information, and vehicle sensor information, are marked as consistent anomalies, a set of candidate scheduling paths is obtained.

[0010] Furthermore, the hierarchical filtering of the candidate scheduling path set includes: In the first layer of screening, the estimated transportation time of each path in the candidate scheduling path set is obtained and compared with a time threshold set according to the historical transportation time distribution. When the estimated transportation time is within the time threshold range, the first candidate path is obtained. In the second layer of screening, the resource carrying capacity of the storage nodes traversed by the first candidate path is determined. The resource carrying capacity is determined based on the real-time inventory quantity and processing capacity. When the resource carrying capacity is greater than or equal to the capacity required by the first candidate path, the second candidate path is obtained. In the third layer of screening, the second candidate path is subjected to a path trajectory consistency determination. The path trajectory consistency determination sets the trajectory range based on the spatial distribution of historical delivery trajectories. When the trajectory of the path is within the trajectory range, the final scheduling path is obtained.

[0011] Furthermore, when generating compensation factors based on customer feedback levels, the following are included: Obtain customer feedback data, including customer rating questionnaires, return rates, and complaint records; According to the preset grading criteria, the customer feedback data is judged into high, medium and low grades. When the customer rating is greater than or equal to the preset upper limit and the return rate and complaint records are within the allowable range, it is judged as high grade. When the customer rating is in the middle range and there is a single abnormality in the return rate or complaint records, it is judged as medium grade. When the customer rating is lower than the preset lower limit or there are multiple abnormalities in the return rate and complaint records, it is judged as low grade. When the customer feedback level is high, the compensation factor is determined to be a positive correction factor; When the customer feedback level is medium, the compensation factor is determined to be a neutral correction factor. When the customer feedback level is low, the compensation factor is determined to be a negative correction factor.

[0012] Furthermore, the logistics operation indicators during the execution process are revised. If the revised logistics operation indicators still show anomalies after being compared with the historical indicator set again, this includes: When the compensation factor is a positive correction factor, the logistics operation indicators are improved and corrected. The improvement and correction include improving the numerical performance of the operation indicators or relaxing the historical threshold range. When the compensation factor is a neutral correction factor, the logistics operation indicators remain unchanged, and the comparison proceeds directly to the next stage. When the compensation factor is a negative correction factor, the logistics operation indicators are downweighted and corrected. The downweighting correction includes reducing the numerical performance of the operation indicators or tightening the historical threshold range. The revised logistics operation indicators were compared again with the historical indicator set; When there are still abnormal situations in the comparison results, emergency path reconstruction and resource reallocation are triggered. The emergency path reconstruction includes regenerating a set of candidate scheduling paths and sequentially filtering them through time thresholds, resource carrying capacity and path trajectory consistency to obtain a new final scheduling path. The resource reallocation includes adjusting the resource carrying capacity among storage nodes.

[0013] Furthermore, when generating tiered early warning information, the following are included: The revised logistics operation indicators are compared with historical indicators, and then classified into Level 1, Level 2, and Level 3 warnings according to the degree of deviation from the largest to the smallest. When a Level 1 warning is generated, the serious anomaly handling procedure is triggered, and the warning information is pushed to the management dashboard and synchronized to all user terminals; When a Level 2 warning is generated, the moderate anomaly handling procedure is triggered, and the warning information is pushed to the operation terminal; When a Level 3 warning is generated, a minor anomaly handling procedure is triggered, the warning information is recorded in the report, and the warning information is incorporated into the subsequent trend analysis of operational indicators.

[0014] Compared with existing technologies, the advantages of this invention are as follows: By standardizing order information, warehouse status, vehicle location, customer feedback, and settlement data, and implementing hierarchical management in a data platform, unified integration of multi-source heterogeneous data is achieved, solving the integration difficulties caused by data dispersion and inconsistent definitions in existing technologies; logistics operation indicators such as transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level are extracted from the hierarchical data and compared with historical indicator sets, enabling accurate identification of operational anomalies in multiple dimensions and reducing the misjudgment rate caused by single indicators or missing data; when anomaly detection results indicate anomalies, vehicle sensor information, weather information, and traffic information are introduced for consistency verification, effectively eliminating interference from accidental or single-point data fluctuations and ensuring the credibility of anomaly judgment; by sequentially performing time threshold, resource carrying capacity, and path trajectory checks on the candidate scheduling path set, the invention achieves unified integration of multi-source heterogeneous data, solving the integration difficulties caused by data dispersion and inconsistent definitions in existing technologies; Consistent hierarchical screening, while meeting timeliness requirements, also considers the processing capacity of warehousing nodes and the rationality of delivery trajectories, ensuring the scientific validity and feasibility of the final scheduling path. During logistics scheduling, compensation factors are generated based on customer feedback levels to correct operational indicators. When the corrected indicators still show anomalies compared with the historical indicator set, emergency path reconstruction and resource reallocation are triggered sequentially, forming an adaptive feedback control mechanism that improves the system's responsiveness and operational stability in emergencies. By generating hierarchical early warning information and classifying it into Level 1, Level 2, and Level 3 warnings according to severity, operators can promptly obtain risk alerts at different levels and take corresponding action, thereby improving the precision of risk management. Finally, the logistics operation indicators, candidate scheduling path sets, corrected operation indicators, and hierarchical early warning information are displayed, enhancing the system's visualization level and user experience.

[0015] On the other hand, this application also provides a data-based logistics management system for applying the above-mentioned data-based logistics management method, including: The data acquisition module acquires business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and performs standardization processing on the business data to obtain standard business data. The hierarchical management module manages the standard business data in the data platform in a hierarchical manner. The hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the hierarchical process, the data is cleaned, denoised, and formatted to obtain hierarchical data. The anomaly detection module extracts logistics operation indicators from the layered data. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. The module then compares these logistics operation indicators with a set of historical indicators to obtain anomaly detection results. The consistency verification module performs consistency verification based on vehicle sensor information, weather information, and traffic information when the anomaly detection result indicates an anomaly. When the consistency verification result is an anomaly, a set of candidate scheduling paths is generated. The path determination module performs hierarchical screening on the candidate scheduling path set, successively filtering by time threshold, resource carrying capacity and path trajectory consistency, to obtain the final scheduling path; The compensation and correction module implements logistics scheduling based on the final scheduling path and generates compensation factors based on customer feedback levels to correct logistics operation indicators during the execution process. When the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation are triggered in sequence. The early warning module generates tiered early warning information based on the revised logistics operation indicators. The tiered early warning information is divided into Level 1, Level 2, and Level 3 early warnings, and triggers corresponding risk alerts and handling procedures. The display module shows the logistics operation indicators, the candidate scheduling path set, the corrected logistics operation indicators, and the hierarchical early warning information.

[0016] It is understandable that the aforementioned logistics management system based on a data platform has the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a logistics management method based on a data middle platform, provided as an embodiment of the present invention; Figure 2 This is a structural block diagram of a logistics management system based on a data middle platform, provided as an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] In some embodiments of this application, see Figure 1 As shown, a logistics management method based on a data platform includes: S100: Acquire business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and standardize the business data to obtain standard business data; S200: Standard business data is managed in layers in the data platform. Layer management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the layering process, cleaning, noise reduction, and format unification are completed to obtain layered data. S300: Extract logistics operation indicators from the hierarchical data. The logistics operation indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation and customer feedback level. The logistics operation indicators are compared with the historical indicator set to obtain anomaly detection results. S400: When the anomaly detection result shows that there is an anomaly, a consistency verification is performed based on vehicle sensor information, weather information and traffic information. When the consistency verification result is an anomaly, a set of candidate scheduling paths is generated. S500: The candidate scheduling path set is screened in a hierarchical manner, successively by time threshold, resource carrying capacity and path trajectory consistency, to obtain the final scheduling path; S600: Implement logistics scheduling based on the final scheduling path, generate compensation factors based on customer feedback levels, and correct logistics operation indicators during the execution process. If the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation will be triggered in sequence. S700: Generates tiered early warning information based on the revised logistics operation indicators. The tiered early warning information is divided into Level 1, Level 2 and Level 3 early warnings, and triggers corresponding risk warnings and handling procedures. S800: Displays logistics operation indicators, candidate scheduling path sets, revised logistics operation indicators, and hierarchical early warning information.

[0020] Specifically, the business data sources include order management systems, warehouse management systems, transportation scheduling systems, and customer service systems. The business data encompasses order information, warehouse status, vehicle location, customer feedback, and settlement data. After standardized processing in terms of format, units, and coding, the business data becomes standard business data. Subsequently, this standard business data undergoes hierarchical management on the data platform. This hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. The operational data layer is the basic layer storing data by field; the detailed data layer is a data layer integrated by business objects and time series; and the data mart layer is a summary layer aggregating data by topic for analysis. This hierarchical management results in layered data. Based on this layered data, logistics operation indicators are extracted. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. Transportation timeliness refers to the time from dispatch to delivery; warehouse turnover refers to the speed of inventory turnover; delivery trajectory deviation refers to the degree of deviation of the vehicle's trajectory from the preset path; and customer feedback level refers to the customer's evaluation of the logistics process. The logistics operation indicators are compared with historical indicator sets to obtain anomaly detection results. When anomalies are detected, consistency verification is performed based on vehicle sensor information, weather information, and traffic information. Vehicle sensor information includes speed, acceleration, and positioning trajectory; weather information includes temperature, humidity, and precipitation; and traffic information includes road congestion and travel time. When the conclusions of this information verification match the anomaly detection results, a set of candidate scheduling paths is generated. This set of candidate scheduling paths undergoes hierarchical screening, sequentially checking for time thresholds, the resource carrying capacity of warehouse nodes, and path trajectory consistency, ultimately resulting in a determined scheduling path. Logistics scheduling is executed based on the final scheduling path, and compensation factors are generated based on customer feedback levels to correct logistics operation indicators during execution. If the corrected logistics operation indicators still show anomalies after comparison with historical indicator sets, emergency path reconstruction and resource reallocation are triggered sequentially. Finally, hierarchical early warning information is generated based on the corrected logistics operation indicators, categorized into Level 1, Level 2, and Level 3 warnings, triggering corresponding risk alerts and handling procedures. Simultaneously, the logistics operation indicators, the set of candidate scheduling paths, the corrected logistics operation indicators, and the hierarchical early warning information are displayed, providing real-time query and decision support for operations personnel and customers. For example, in a city-wide delivery scenario, the system first collects order information, warehouse status, vehicle location, customer feedback, and settlement data. It then unifies data from different sources into the same time format, unit of measurement, and encoding method to obtain standard business data. Subsequently, the system manages this standard business data hierarchically within the data platform: the operational data layer saves the original records one by one; the detailed data layer organizes them according to order and time sequence; and the data mart layer aggregates them into indicators such as transportation timeliness, warehouse turnover, delivery trajectory, and customer feedback.During operation, the system extracts the transportation timeliness and warehousing turnover of this batch of orders. Comparison with historical statistics reveals that the transportation timeliness is slow. Based on weather and vehicle sensor data, it detects rainfall along the route and a decrease in average vehicle speed, consistent with the abnormal transportation timeliness. Therefore, the system generates new candidate scheduling routes. After three layers of screening based on transportation time, warehousing resources, and trajectory consistency, the final scheduling route is obtained and logistics scheduling is implemented according to this route. During scheduling, the system generates compensation factors based on customer feedback levels. For example, if customer feedback is low, the system corrects the indicators to a negative level and strengthens scheduling monitoring. If the corrected operational indicators are still abnormal when compared with historical results, the system triggers emergency route reconstruction and reallocation of warehousing resources. Finally, the system generates tiered early warning information based on the degree of deviation and displays operational indicators, scheduling routes, and early warning results on a dashboard.

[0021] Understandably, by standardizing and hierarchically managing multi-source business data through a data platform, order information, warehouse status, vehicle location, customer feedback, and settlement data can be integrated and processed within a unified framework, thus ensuring the accuracy and comparability of operational indicator extraction. Based on this, consistency verification is performed using vehicle sensor information, weather information, and traffic information, avoiding misjudgments caused by single indicator anomalies and improving the reliability of anomaly detection. By tiered screening of candidate scheduling paths, a balance can be achieved between transportation timeliness, warehouse resources, and trajectory rationality, ensuring the executability of path decisions. According to the compensation factor mechanism, customer feedback levels directly affect the indicator correction process, achieving dynamic linkage between user experience and scheduling decisions. Furthermore, when corrected operational indicators still show anomalies, emergency path reconstruction and resource reallocation are triggered sequentially, enabling the system to have self-adjustment capabilities. Finally, the generation and display of tiered early warning information provide different levels of risk alerts and decision support for management, operations personnel, and customers. Therefore, this application can achieve real-time perception of operational status, anomaly identification, path optimization, and risk warning in complex and ever-changing logistics environments, improving the stability and response efficiency of logistics management.

[0022] In some embodiments of this application, the process of standardizing business data to obtain standard business data includes: Standardize the format of the time field in business data and convert time records from different sources into a unified timestamp format. Unify the numerical fields in business data to the International System of Units (SI); By unifying the text fields of business data to UTF-8 encoding, standard business data is obtained.

[0023] Specifically, in the process of standardizing business data, firstly, for the time field, all time records from different business systems are converted into a unified timestamp format to eliminate alignment deviations caused by differences in recording methods; secondly, for the numerical fields, values ​​involving weight, volume, distance, and amount are unified into the International System of Units (SI) to ensure that data from different sources have a consistent measurement standard; finally, for the text fields, text involving order descriptions, customer information, and warehouse records are unified into UTF-8 encoding to avoid garbled characters or missing information due to inconsistent character sets.

[0024] In some embodiments of this application, when standard business data is managed hierarchically in a data platform, the following are included: When establishing the operational data layer, standard business data is stored one by one according to field dimensions, and missing value filling, abnormal record removal and duplicate record deduplication are performed during the storage process to obtain operational layer data; When establishing the detailed data layer, the operational layer data is integrated according to business objects and time series. During the integration process, noise reduction and format unification are performed on cross-system records of the same business object to generate detailed data. In the process of establishing the data mart layer, detailed data is used as the source, and aggregated according to transportation timeliness, warehouse turnover, delivery trajectory and customer feedback. During the aggregation process, the standardization of indicators from different sources is adjusted to form hierarchical data.

[0025] Specifically, when establishing the operational data layer, standardized business data is written to the storage table row by row according to fields such as order number, warehouse number, vehicle number, and customer number. During the writing process, missing field values ​​are supplemented by records of the same business object in adjacent time windows. Obviously abnormal field values ​​are removed by comparing with historical distribution ranges. Duplicate records are identified by primary key indexes and deduplicated, thus obtaining operational layer data that can be directly accessed. When establishing the detailed data layer, based on the operational layer data, it is integrated according to the operation of individual business objects in a continuous time series. Records of the same business object across systems and stages are unified by unique identifiers. During the integration process, format uniformity and noise removal are performed on record format differences, thereby generating detailed data that is oriented towards a single object and has a complete time sequence. When establishing the data mart layer, detailed data is used as input, and aggregated statistics are performed around four thematic dimensions: transportation timeliness, warehouse turnover, delivery trajectory, and customer feedback. During the aggregation process, the statistical standards of indicators from different business systems are consistent to ensure that the same indicators are consistent in statistical standards, and finally, layered data that can be used for operational analysis and indicator extraction is formed.

[0026] In some embodiments of this application, the process of extracting logistics operation indicators from hierarchical data and comparing them with a historical indicator set includes: The transportation timeliness is compared with the historical average. When the difference exceeds the time threshold, the transportation timeliness comparison result is marked as abnormal. The time threshold is obtained based on the statistical distribution range of historical transportation timeliness data. The warehouse turnover rate is compared with the historical average. When the difference exceeds the turnover threshold, the warehouse turnover comparison result is marked as abnormal. The turnover threshold is determined based on the fluctuation range of the historical warehouse turnover rate. The delivery trajectory deviation is compared with the historical trajectory range. When the delivery trajectory deviation is not within the historical trajectory range, the delivery trajectory comparison result is marked as abnormal. The historical trajectory range is set by the spatial coverage range of multi-day delivery trajectory data. The customer feedback level is compared with the historical distribution range. When the customer feedback level is low and below the lower limit of the historical distribution range, the customer feedback comparison result is marked as abnormal. The historical distribution range is divided into upper limit range, middle range and lower limit range based on the statistical results of customer feedback level in several historical periods. An anomaly detection result is generated when any of the comparison results of transportation timeliness, warehousing turnover, delivery trajectory, and customer feedback is marked as abnormal.

[0027] Specifically, when extracting logistics operation indicators from layered data and comparing them with historical indicator sets, the following steps are taken: First, transportation timeliness is assessed. This involves creating a transportation timeliness time series by accessing transportation timeliness data from multiple historical periods. Statistical distribution analysis is then performed on this time series to extract the mean, standard deviation, and quantile intervals of transportation timeliness. Based on the statistical results, a normal fluctuation range for transportation timeliness is determined, and the boundary values ​​of this range are used as the transportation timeliness threshold. When the difference between the actual transportation timeliness and the mean transportation timeliness exceeds the transportation timeliness threshold, the transportation timeliness comparison result is judged as abnormal. Second, warehouse turnover is assessed. This involves creating a warehouse turnover sequence by accessing warehouse turnover data from multiple historical periods. The fluctuation range of the warehouse turnover sequence is divided, and a warehouse turnover threshold is determined by combining the average cycle and the upper and lower fluctuation ranges of warehouse turnover. When the actual warehouse turnover result exceeds the warehouse turnover threshold... During the assessment, the warehouse turnover comparison results are identified as abnormal. Further analysis of delivery trajectories is conducted by calling multi-day delivery trajectory data and overlaying it on a data platform to generate a distribution of delivery trajectory coverage areas. The range of historical delivery trajectories is determined based on their spatial density and coverage boundaries. When the actual delivery trajectory deviates from the historical trajectory range, the delivery trajectory comparison result is identified as abnormal. Regarding customer feedback levels, customer feedback data is generated by combining customer rating questionnaires, return rates, and complaint records, and compared with customer feedback distribution intervals formed over several historical periods. When the customer feedback level is low and below the lower limit of the customer feedback distribution interval, the customer feedback comparison result is identified as abnormal. An anomaly detection result is generated when any of the transportation timeliness comparison results, warehouse turnover comparison results, delivery trajectory comparison results, or customer feedback comparison results are identified as abnormal.

[0028] In some embodiments of this application, consistency verification is performed when an anomaly detection result indicates the presence of an anomaly, including: Based on weather information, real-time temperature, humidity, and precipitation type are extracted, and the presence of meteorological conditions that could cause transportation delays is determined based on temperature fluctuations, humidity trends, and precipitation conditions, thus obtaining the verification results of the weather information. When both the verification results of the weather information and the abnormal comparison results of transportation timeliness indicate that there is a transportation delay, the transportation timeliness comparison results are marked as consistent anomalies. Based on traffic information, the road congestion index, road segment travel time and historical average are extracted, and the presence of traffic conditions that lead to an extension of the cargo turnover cycle is determined based on road congestion and travel time, thus obtaining the verification results of traffic information; when both the verification results of traffic information and the abnormal comparison results of warehouse turnover determine that the cargo turnover cycle is extended, the warehouse turnover comparison results are marked as consistent anomalies. Based on vehicle sensor information, the vehicle's speed, acceleration, and positioning trajectory are extracted. The presence of path deviation is determined based on speed fluctuations, acceleration changes, and positioning trajectory offsets, thus obtaining the verification results of the vehicle sensor information. When both the verification results of the vehicle sensor information and the abnormal comparison results of the delivery trajectory deviation indicate that there is a path deviation, the delivery trajectory deviation comparison results are marked as consistent anomalies. When the verification results of two or more types of information, including weather information, traffic information, and vehicle sensor information, are marked as consistent anomalies, a set of candidate scheduling paths is obtained.

[0029] Specifically, real-time temperature, humidity, and precipitation type are extracted from weather information. Temperature and humidity sequences are obtained through unified time-step resampling and outlier removal. Precipitation types are mapped to fixed categories. Temperature fluctuation amplitude is determined by the difference between the average value within a short-term time window and the long-term baseline. The long-term baseline is the center position of historical temperature data from multiple periods within the same region. The temperature fluctuation threshold is set based on the high quantile or upper bound of the dispersion of historical temperature distribution. Humidity change trends are determined by the monotonicity and cumulative change within a short-term time window. The trend threshold is set based on the upper bound of humidity change distribution during historical stable periods. Precipitation conditions are determined by whether the real-time precipitation type and intensity reach the intensity threshold of the historical distribution. If the temperature fluctuation amplitude exceeds the temperature fluctuation threshold, or the humidity change trend exceeds the trend threshold, or unfavorable precipitation reaches the intensity threshold and its duration exceeds the minimum duration threshold, the weather information verification result indicates a transportation delay. When both the weather information verification result and the anomaly comparison result for transportation timeliness indicate a transportation delay, the conclusions are the same, and the transportation timeliness comparison result is marked as a consistent anomaly. The verification process for traffic information is as follows: Road congestion index and road segment travel time are extracted and mapped by road segment, direction, and time period. The congestion threshold is derived from the high quantile or upper bound of the dispersion of historical congestion index distributions for the same road segment and time period. The travel time threshold is derived from the high quantile or upper bound of the dispersion of historical travel time distributions for the same road segment and time period. When the congestion index continuously exceeds the congestion threshold and the travel time continuously exceeds the travel time threshold, and the duration exceeds the minimum duration threshold, the traffic information verification result is determined to be a traffic condition leading to an extended freight turnover cycle. When the traffic information verification result determines that a traffic condition leading to an extended freight turnover cycle exists, and the abnormal comparison result of warehouse turnover also determines that the freight turnover cycle is extended, the two conclusions are the same, and the warehouse turnover comparison result is marked as consistent anomaly.The vehicle sensor information verification process is as follows: extract vehicle speed, acceleration, and positioning trajectory, and perform time synchronization, outlier removal, and map matching. The speed fluctuation range is determined by the speed dispersion index within a short window, with the threshold derived from the upper bound of the dispersion distribution of historical normal driving samples under similar road conditions and time periods. Acceleration change is determined by the cumulative change in acceleration difference between adjacent time points, with the threshold derived from the upper bound of the cumulative change distribution of historical normal driving samples. Positioning trajectory deviation is determined by the shortest distance sequence from the actual trajectory to the preset path, with the threshold derived from the spatial boundary and tolerance zone generated by the historical delivery trajectory range. When at least two of the following conditions are met—speed fluctuation range exceeding the dispersion threshold, acceleration change exceeding the change threshold, and positioning trajectory deviation exceeding the distance threshold—and the duration exceeds the shortest duration threshold, the vehicle sensor information verification result is determined to have path deviation. When the vehicle sensor information verification result determines that there is path deviation, and the abnormal comparison result of the delivery trajectory deviation also determines that there is path deviation, the two conclusions are the same, and the delivery trajectory deviation comparison result is marked as consistent anomaly. Finally, a set of candidate scheduling paths is generated when at least two types of verification results from weather information, traffic information, and vehicle sensor information are marked as consistent anomalies.

[0030] For example, in a delivery scenario, layered data comparison reveals abnormal delivery times. Weather information verification shows that the real-time temperature in the area has risen by more than 5°C compared to the historical average over the past two hours, humidity has continuously decreased by more than 10%, and heavy rainfall has been detected, reaching the upper limit of the historical rainfall intensity threshold. Therefore, the weather information verification result indicates a delivery delay. Simultaneously, traffic information verification shows that the congestion index of the main roads along the delivery route is continuously higher than the historical high percentile threshold, travel time exceeds 50% of the historical average, and the duration exceeds 30 minutes. Therefore, the traffic information verification result indicates an extended cargo turnover cycle. Furthermore, vehicle sensor information indicates that vehicle speed fluctuations exceed the historical dispersion threshold, acceleration changes frequently, and the positioning trajectory deviates from the historical path range multiple times. The vehicle sensor information verification result indicates a path deviation. Since the verification results of weather and traffic information are consistent with the comparison results of abnormal transportation timeliness and abnormal warehouse turnover, and the verification results of vehicle sensor information are consistent with the comparison results of abnormal delivery trajectory deviation, at least three types of information are marked as consistent anomalies. The system then generates a set of candidate scheduling paths and searches for alternative paths that can avoid rainstorms and congested road sections in the candidate path set for subsequent screening and scheduling.

[0031] In some embodiments of this application, the hierarchical filtering of the candidate scheduling path set includes: In the first layer of screening, the estimated transportation time of each path in the candidate scheduling path set is obtained and compared with a time threshold set according to the historical transportation time distribution. When the estimated transportation time is within the time threshold range, the first candidate path is obtained. In the second layer of screening, the resource carrying capacity of the storage nodes through which the first candidate path passes is determined. The resource carrying capacity is determined based on the real-time inventory quantity and processing capacity. When the resource carrying capacity is greater than or equal to the capacity required by the first candidate path, the second candidate path is obtained. In the third layer of screening, the consistency of the path trajectory is determined for the second candidate path. The consistency determination of the path trajectory is based on the spatial distribution of historical delivery trajectories, and the trajectory range is set. When the trajectory of the path is within the trajectory range, the final scheduling path is obtained.

[0032] Specifically, in the first-level screening, the estimated transportation time of each candidate scheduling path is obtained. The estimated transportation time is converted according to the combination rules of path segment mileage, historical average speed in the same period, and real-time traffic weight factor. The historical average speed comes from the speed distribution statistics of the same road segments in the same period of multiple cycles, and the real-time traffic weight factor comes from the current congestion level segment table. The time threshold range comes from the transportation time distribution of the same path in multiple cycles. The normal fluctuation range is divided according to the central tendency and dispersion, and the boundary of the range is used as the threshold. When the estimated transportation time falls within the time threshold range, the path is marked as the first candidate path. Subsequently, in the second-level screening, the resource carrying capacity of each warehousing node traversed by the first candidate route is determined. The resource carrying capacity is expressed using a unified capacity unit. The unified capacity unit is achieved by converting the number of items, volume, and weight into the same measurement caliber according to the loading conversion factor. The loading conversion factor is derived from the statistics of historical loading and inbound / outbound operation parameters. The resource carrying capacity is jointly limited by three parts: available inventory, unit time processing capacity, and platform throughput capacity. Available inventory refers to the standard capacity that can be shipped out or transferred at present. Unit time processing capacity refers to the maximum standard capacity that loading, unloading, and sorting can complete within a specified time window. Platform throughput capacity refers to the maximum standard capacity corresponding to the number of batches of vehicles that can be completed within a specified time window. The resource carrying capacity of a node is determined according to the minimum constraint principle, that is, the smallest value among the three parts is taken as the final resource carrying capacity. The required capacity of the first candidate route is calculated and summarized by converting the planned cargo volume of the route within the planned arrival time window of each node according to the unified capacity unit. When the resource carrying capacity of each node in the corresponding time window is greater than or equal to the required capacity, the route is marked as the second candidate route. Finally, in the third layer of screening, the consistency of the path trajectory of the second candidate path is determined. The spatial distribution of historical delivery trajectories is generated by superimposing delivery trajectories from multiple days to form a coverage area, and tolerance bands are set on both sides of the preset path. The width of the tolerance band comes from the upper quantile value of the historical deviation distance distribution. The consistency of the path trajectory is determined by comparing the spatial relationship between the geographic coordinate sequence of the second candidate path and the coverage area and the tolerance band. When the geographic coordinate sequence is entirely within the coverage area or the tolerance band, it means that all sampling points fall within the coverage area or the tolerance band. When the proportion of sampling points in the geographic coordinate sequence that are within the coverage area or the tolerance band is not less than the consistency proportion threshold, it is also considered to meet the consistency condition. The consistency proportion threshold comes from the spatial landing point proportion distribution of historical compliant paths.

[0033] In some embodiments of this application, generating a compensation factor based on customer feedback levels includes: Obtain customer feedback data, which includes customer rating questionnaires, return rates, and complaint records; Customer feedback data is classified into high, medium, and low levels according to the preset grading criteria. A customer rating is classified as high if it is greater than or equal to the preset upper limit and the return rate and complaint records are within the allowable range. A customer rating is classified as medium if it is in the middle range and there is a single abnormality in the return rate or complaint records. A customer rating is classified as low if it is lower than the preset lower limit or there are multiple abnormalities in the return rate and complaint records. When the customer feedback level is high, the compensation factor is determined to be a positive correction factor; When the customer feedback level is medium, the compensation factor is determined to be a neutral correction factor. When the customer feedback level is low, the compensation factor is determined to be a negative correction factor.

[0034] Specifically, customer feedback data consists of customer questionnaire ratings after order fulfillment, return event records, and after-sales complaint records. Before entering the processing stage, all three types of data are linked by order number and timestamps are aligned to ensure consistency of source and validity of evaluation. The rating criteria were established during the system deployment phase by statistically analyzing customer feedback data across multiple historical periods. Specifically: First, using a six-month or one-year period as the statistical window, frequency distribution tables were generated for questionnaire scores, return rates, and complaint records. Then, intervals were defined based on the central tendency and quantile of each indicator in the distribution tables. Customer scores were divided into three intervals according to the median and quantiles of the distribution: scores of 80 or higher corresponded to the high-score interval; scores of 60 to 79 (inclusive) corresponded to the middle interval; and scores below 60 corresponded to the low-score interval. Return rates were divided into three intervals based on the quantile of cumulative frequency. Historically, 75% of orders had a return rate below 5%, therefore, a return rate not exceeding 5% was set as the upper limit, 6% to 15% as the middle interval, and above 15% as the lower limit. Complaint records were directly set based on the number of incidents: no complaints were set as the upper limit, one complaint as the middle interval, and two or more complaints as the lower limit. This method ensures that the interval boundaries are derived from long-term statistical patterns, avoiding subjective designation. Based on the above three categories of indicators, high-level customer feedback is defined as a customer rating falling into the high-score range with both return rate and complaint records at the upper limit. Medium-level customer feedback is defined as a customer rating falling into the middle range with either the return rate or complaint records at the middle range while the other is not lower than the middle range. Low-level customer feedback is defined as a customer rating falling into the low-score range or both the return rate and complaint records falling into the lower limit. After the level determination, high-level corresponds to a positive correction factor, medium-level corresponds to a neutral correction factor, and low-level corresponds to a negative correction factor. These correction factors are used in subsequent adjustments to logistics operation indicators to relax, maintain, or tighten the determination conditions, respectively, thereby achieving differentiated driving of operational indicator adjustments based on customer feedback information.

[0035] In some embodiments of this application, when logistics operation indicators during the execution process are corrected, and the corrected logistics operation indicators still show abnormalities after being compared again with the historical indicator set, the following steps are taken: When the compensation factor is a positive correction factor, the logistics operation indicators are improved and corrected. The improvement and correction include improving the numerical performance of the operation indicators or widening the historical threshold range. When the compensation factor is a neutral correction factor, the logistics operation indicators remain unchanged, and the comparison proceeds directly to the next stage. When the compensation factor is a negative correction factor, the logistics operation indicators are downweighted and corrected. Downweighting correction includes reducing the numerical performance of the operation indicators or tightening the historical threshold range. The revised logistics operation indicators were compared again with the historical indicator set; When there are still abnormal situations in the comparison results, emergency path reconstruction and resource reallocation are triggered. Emergency path reconstruction includes regenerating the candidate scheduling path set and sequentially filtering by time threshold, resource carrying capacity and path trajectory consistency to obtain a new final scheduling path. Resource reallocation includes adjusting the resource carrying capacity among storage nodes.

[0036] Specifically, when the compensation factor is a positive correction factor, the logistics operation indicators are improved. This improvement refers to appropriately widening the allowable boundaries of deviations in transportation timeliness, warehousing turnover, and delivery routes without changing the meaning of the indicators themselves. For example, the upper limit of allowable transportation timeliness is extended to a level that has maintained normal operation for most periods in historical statistics, and the lower limit of warehousing turnover rate is adjusted to the lower end of the historical fluctuation range, thus reflecting the larger tolerance range represented by positive customer feedback. When the compensation factor is a neutral correction factor, the original threshold conditions remain unchanged, and the operation indicators are directly compared with the historical indicator set again. When the compensation factor is a negative correction factor, the logistics operation indicators are downweighted. This downweighting refers to tightening the threshold range during the judgment process. For example, the upper limit of allowable transportation timeliness is shortened to near the historical average, and the lower limit of warehousing turnover rate is raised to the level of a stable period, reflecting the strict requirements corresponding to negative customer feedback. After the correction is completed, each corrected logistics operation indicator is compared with the historical indicator set to determine whether it falls within the corresponding threshold range. If any indicator in the re-comparison results is still deemed abnormal, emergency path reconstruction and resource reallocation are triggered. Emergency path reconstruction involves regenerating a set of candidate scheduling paths and sequentially filtering them according to time thresholds, resource carrying capacity, and path trajectory consistency to obtain a new final scheduling path. Resource reallocation involves redistributing standard capacity among nodes based on real-time inventory levels, unit-time processing capacity, and platform access capacity to ensure the smooth execution of the new final scheduling path. For example, if a large number of customer returns and complaints are reported, the system generates a negative correction factor, tightening the upper limit of transportation timeliness to a more stringent range. The original path is eliminated because it exceeds the tightened timeliness threshold. The system regenerates candidate paths and ultimately selects a path that passes through a backup warehouse node. At the same time, through resource reallocation, some goods are transferred to the backup warehouse node in advance, thereby ensuring that the overall logistics scheduling can still be completed under stricter constraints.

[0037] In some embodiments of this application, generating graded early warning information includes: The revised logistics operation indicators are compared with historical indicators, and then classified into Level 1, Level 2, and Level 3 warnings according to the degree of deviation from the largest to the smallest. When a Level 1 warning is generated, the serious anomaly handling procedure is triggered, and the warning information is pushed to the management dashboard and synchronized to all user terminals; When a Level 2 warning is generated, the moderate anomaly handling procedure is triggered, and the warning information is pushed to the operation terminal; When a Level 3 warning is generated, a minor anomaly handling procedure is triggered, the warning information is recorded in the report, and the warning information is incorporated into the subsequent trend analysis of operational indicators.

[0038] Specifically, when generating tiered early warning information, the revised logistics operation indicators are first compared item by item with the historical indicator set to determine whether each indicator deviates and to rank the degree of deviation. When the transportation timeliness exceeds the strict interval boundary compared to the historical distribution, the warehouse turnover rate is lower than the historical fluctuation lower limit, the delivery trajectory deviation exceeds the coverage range of the historical trajectory, or the customer feedback level falls into a low level and is lower than the historical distribution interval lower limit, it can be identified as a serious anomaly, generating a Level 1 early warning and triggering the serious anomaly handling process. This process includes immediately pushing the early warning information to the management dashboard and synchronizing it to all user terminals so that management and all users can receive risk alerts as soon as possible. When the transportation timeliness is close to but does not exceed the warning boundary of the historical distribution, and the warehouse turnover rate is fluctuating... If the delivery trajectory deviation is significant but still within the tolerance range between the lower limit and the middle range, or if the customer feedback level falls into the medium level range, it is considered a moderate anomaly, generating a level 2 warning and triggering the moderate anomaly handling process. The warning information is pushed to the operation terminal so that operators can adjust the scheduling strategy in a timely manner. When the transportation timeliness is on the edge of the historical fluctuation range, the warehouse turnover rate is slightly lower than the historical average but higher than the lower limit, the delivery trajectory deviation is on the edge of the tolerance range, or the customer feedback level is low but still within the historical distribution range, it is considered a mild anomaly, generating a level 3 warning and triggering the mild anomaly handling process. The warning information is recorded in the report and used as input data for subsequent operational indicator trend analysis to identify the accumulation trend of potential risks in subsequent analysis.

[0039] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides a logistics management system based on a data middle platform, including: The data acquisition module acquires business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and performs standardization processing on the business data to obtain standard business data. The hierarchical management module manages standard business data in the data platform in layers. Hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the layering process, cleaning, noise reduction, and format unification are completed to obtain layered data. The anomaly detection module extracts logistics operation indicators from the layered data. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. The module then compares these logistics operation indicators with historical indicator sets to obtain anomaly detection results. The consistency verification module performs consistency verification based on vehicle sensor information, weather information, and traffic information when the anomaly detection result shows an anomaly. When the consistency verification result is an anomaly, a set of candidate scheduling paths is generated. The path determination module performs hierarchical screening of the candidate scheduling path set, successively filtering based on time threshold, resource carrying capacity, and path trajectory consistency, to obtain the final scheduling path; The compensation and correction module implements logistics scheduling based on the final scheduling path and generates compensation factors based on customer feedback levels to correct logistics operation indicators during the execution process. When the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation are triggered in sequence. The early warning module generates tiered early warning information based on the revised logistics operation indicators. The tiered early warning information is divided into Level 1, Level 2 and Level 3 early warnings, and triggers corresponding risk alerts and handling procedures. The display module shows logistics operation indicators, candidate scheduling path sets, corrected logistics operation indicators, and hierarchical early warning information.

[0040] Understandably, this invention achieves centralized access and standardized processing of multi-source business data through a data acquisition module, ensuring the uniformity of data from different business systems in terms of time, value, and encoding. A hierarchical management module establishes operational data, detailed data, and data mart layers, enabling step-by-step processing from raw data to aggregated indicators, ensuring data integrity and consistency in storage, integration, and aggregation. An anomaly detection module extracts four operational indicators—transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level—from the hierarchical data and compares them with historical indicator sets, enabling timely detection of deviations and anomalies during operation. When an anomaly is detected, a consistency verification module performs cross-validation using vehicle sensor information, weather information, and traffic information, avoiding misjudgments from a single data source. Therefore, a candidate scheduling path set is generated only when the verification results of multiple types of information are consistent, improving the reliability of anomaly detection. Reliability; A multi-level screening mechanism is established through the path determination module, which considers time thresholds, resource carrying capacity of storage nodes, and path trajectory consistency in turn, so that the final determined scheduling path can meet the operational requirements in terms of timeliness, resource and spatial distribution; The compensation and correction module introduces compensation factors based on customer feedback levels in the logistics scheduling process to realize dynamic correction of operational indicators, and triggers emergency path reconstruction and resource reallocation in turn when the correction is ineffective, ensuring the flexibility and stability of the scheduling process; The early warning module generates graded early warning information and establishes corresponding handling procedures to realize graded response from mild, moderate to severe anomalies; The display module visualizes operational indicators, candidate paths, correction results and early warning information, supporting real-time query and decision-making by operators and customers, thus realizing a closed-loop management of data integration, accurate anomaly identification, dynamic scheduling optimization and risk graded early warning.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A logistics management method based on a data middle platform, characterized in that, include: Acquire business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and standardize the business data to obtain standard business data; The standard business data is managed hierarchically in the data platform. The hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the hierarchical process, cleaning, noise reduction, and format unification are completed to obtain hierarchical data. Logistics operation indicators are extracted from the hierarchical data. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. The logistics operation indicators are then compared with a set of historical indicators to obtain anomaly detection results. When the anomaly detection result indicates an anomaly, a consistency verification is performed based on vehicle sensor information, weather information, and traffic information. When the consistency verification result is anomaly, a set of candidate scheduling paths is generated. The candidate scheduling path set is subjected to hierarchical screening, which is carried out by time threshold, resource carrying capacity and path trajectory consistency in turn, to obtain the final scheduling path; Logistics scheduling is implemented according to the final scheduling path, and a compensation factor is generated based on the customer feedback level to correct the logistics operation indicators during the execution process. When the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation are triggered in sequence. Based on the revised logistics operation indicators, a tiered early warning information is generated, which is divided into Level 1, Level 2 and Level 3 early warnings, and triggers corresponding risk warnings and handling procedures. The logistics operation indicators, candidate scheduling path set, corrected logistics operation indicators, and hierarchical early warning information are displayed.

2. The logistics management method based on a data platform according to claim 1, characterized in that, When standardizing the business data to obtain standard business data, the process includes: The time field of the business data is formatted uniformly, and time records from different sources are converted into a unified timestamp format. The numerical fields in the aforementioned business data shall be standardized to the International System of Units (SI). The text fields of the business data are unified to UTF-8 encoding to obtain standard business data.

3. The logistics management method based on a data platform according to claim 2, characterized in that, When managing the standard business data in a hierarchical manner within the data platform, the following is included: When establishing the operational data layer, the standard business data is stored one by one according to the field dimension, and missing value supplementation, abnormal record removal and duplicate record deduplication are performed during the storage process to obtain the operational layer data; When establishing the detailed data layer, the operation layer data is integrated according to business objects and time series. During the integration process, noise reduction and format unification are performed on cross-system records of the same business object to generate detailed data. In the process of establishing the data mart layer, the detailed data is used as the source and aggregated according to transportation timeliness, warehouse turnover, delivery trajectory and customer feedback. During the aggregation process, the standardization of indicators from different sources is adjusted to form hierarchical data.

4. The logistics management method based on a data middle platform according to claim 3, characterized in that, When extracting logistics operation indicators from the hierarchical data and comparing them with a set of historical indicators, the following steps are included: The transportation timeliness is compared with the historical average. When the difference exceeds the time threshold, the transportation timeliness comparison result is marked as abnormal. The time threshold is obtained based on the statistical distribution range of historical transportation timeliness data. The warehouse turnover rate is compared with the historical average. When the difference exceeds the turnover threshold, the warehouse turnover comparison result is marked as abnormal. The turnover threshold is obtained based on the fluctuation range of the historical warehouse turnover rate. The delivery trajectory deviation is compared with the historical trajectory range. When the delivery trajectory deviation is not within the historical trajectory range, the delivery trajectory comparison result is marked as abnormal. The historical trajectory range is set by the spatial coverage range of multi-day delivery trajectory data. The customer feedback level is compared with the historical distribution range. When the customer feedback level is low and below the lower limit of the historical distribution range, the customer feedback comparison result is marked as abnormal. The historical distribution range is divided into upper limit range, middle range and lower limit range based on the statistical results of customer feedback level in several historical periods. An anomaly detection result is generated when any of the comparison results of transportation timeliness, warehousing turnover, delivery trajectory, and customer feedback is marked as abnormal.

5. The logistics management method based on a data middle platform according to claim 4, characterized in that, When the anomaly detection result indicates the presence of an anomaly, a consistency verification is performed, including: Based on weather information, real-time temperature, humidity, and precipitation type are extracted, and the presence of meteorological conditions that could cause transportation delays is determined based on temperature fluctuations, humidity trends, and precipitation conditions, thus obtaining the verification results of the weather information. When both the verification results of the weather information and the abnormal comparison results of transportation timeliness indicate that there is a transportation delay, the transportation timeliness comparison results are marked as consistent anomalies. Based on traffic information, the road congestion index, road segment travel time and historical average are extracted, and the presence of traffic conditions that lead to an extension of the cargo turnover cycle is determined based on road congestion and travel time, thus obtaining the verification results of traffic information; when both the verification results of traffic information and the abnormal comparison results of warehouse turnover determine that the cargo turnover cycle is extended, the warehouse turnover comparison results are marked as consistent anomalies. Based on vehicle sensor information, the vehicle's speed, acceleration, and positioning trajectory are extracted. The presence of path deviation is determined based on speed fluctuations, acceleration changes, and positioning trajectory offsets, thus obtaining the verification results of the vehicle sensor information. When both the verification results of the vehicle sensor information and the abnormal comparison results of the delivery trajectory deviation indicate that there is a path deviation, the delivery trajectory deviation comparison results are marked as consistent anomalies. When the verification results of two or more types of information, including weather information, traffic information, and vehicle sensor information, are marked as consistent anomalies, a set of candidate scheduling paths is obtained.

6. The logistics management method based on a data platform according to claim 5, characterized in that, When performing hierarchical filtering on the candidate scheduling path set, the following is included: In the first layer of screening, the estimated transportation time of each path in the candidate scheduling path set is obtained and compared with a time threshold set according to the historical transportation time distribution. When the estimated transportation time is within the time threshold range, the first candidate path is obtained. In the second layer of screening, the resource carrying capacity of the storage nodes traversed by the first candidate path is determined. The resource carrying capacity is determined based on the real-time inventory quantity and processing capacity. When the resource carrying capacity is greater than or equal to the capacity required by the first candidate path, the second candidate path is obtained. In the third layer of screening, the second candidate path is subjected to a path trajectory consistency determination. The path trajectory consistency determination sets the trajectory range based on the spatial distribution of historical delivery trajectories. When the trajectory of the path is within the trajectory range, the final scheduling path is obtained.

7. The logistics management method based on a data platform according to claim 6, characterized in that, When generating compensation factors based on customer feedback levels, the following are included: Obtain customer feedback data, including customer rating questionnaires, return rates, and complaint records; According to the preset grading criteria, the customer feedback data is judged into high, medium and low grades. When the customer rating is greater than or equal to the preset upper limit and the return rate and complaint records are within the allowable range, it is judged as high grade. When the customer rating is in the middle range and there is a single abnormality in the return rate or complaint records, it is judged as medium grade. When the customer rating is lower than the preset lower limit or there are multiple abnormalities in the return rate and complaint records, it is judged as low grade. When the customer feedback level is high, the compensation factor is determined to be a positive correction factor; When the customer feedback level is medium, the compensation factor is determined to be a neutral correction factor. When the customer feedback level is low, the compensation factor is determined to be a negative correction factor.

8. The logistics management method based on a data platform according to claim 7, characterized in that, If the logistics operation indicators during the execution process are corrected, and the corrected logistics operation indicators still show anomalies after being compared with the historical indicator set again, including: When the compensation factor is a positive correction factor, the logistics operation indicators are improved and corrected. The improvement and correction include improving the numerical performance of the operation indicators or relaxing the historical threshold range. When the compensation factor is a neutral correction factor, the logistics operation indicators remain unchanged, and the comparison proceeds directly to the next stage. When the compensation factor is a negative correction factor, the logistics operation indicators are downweighted and corrected. The downweighting correction includes reducing the numerical performance of the operation indicators or tightening the historical threshold range. The revised logistics operation indicators were compared again with the historical indicator set; When there are still abnormal situations in the comparison results, emergency path reconstruction and resource reallocation are triggered. The emergency path reconstruction includes regenerating a set of candidate scheduling paths and sequentially filtering them through time thresholds, resource carrying capacity and path trajectory consistency to obtain a new final scheduling path. The resource reallocation includes adjusting the resource carrying capacity among storage nodes.

9. The logistics management method based on a data platform according to claim 8, characterized in that, When generating tiered early warning information, the following are included: The revised logistics operation indicators are compared with historical indicators, and then classified into Level 1, Level 2, and Level 3 warnings according to the degree of deviation from the largest to the smallest. When a Level 1 warning is generated, the serious anomaly handling procedure is triggered, and the warning information is pushed to the management dashboard and synchronized to all user terminals; When a Level 2 warning is generated, the moderate anomaly handling procedure is triggered, and the warning information is pushed to the operation terminal; When a Level 3 warning is generated, a minor anomaly handling procedure is triggered, the warning information is recorded in the report, and the warning information is incorporated into the subsequent trend analysis of operational indicators.

10. A logistics management system based on a data platform, used to apply the logistics management method based on a data platform as described in any one of claims 1-9, characterized in that, include: The data acquisition module acquires business data, including order information, warehouse status, vehicle location, customer feedback and settlement data, and performs standardization processing on the business data to obtain standard business data. The hierarchical management module manages the standard business data in the data platform in a hierarchical manner. The hierarchical management includes establishing an operational data layer, a detailed data layer, and a data mart layer. During the hierarchical process, the data is cleaned, denoised, and formatted to obtain hierarchical data. The anomaly detection module extracts logistics operation indicators from the layered data. These indicators include transportation timeliness, warehouse turnover, delivery trajectory deviation, and customer feedback level. The module then compares these logistics operation indicators with a set of historical indicators to obtain anomaly detection results. The consistency verification module performs consistency verification based on vehicle sensor information, weather information, and traffic information when the anomaly detection result indicates an anomaly. When the consistency verification result is an anomaly, a set of candidate scheduling paths is generated. The path determination module performs hierarchical screening on the candidate scheduling path set, successively filtering by time threshold, resource carrying capacity and path trajectory consistency, to obtain the final scheduling path; The compensation and correction module implements logistics scheduling based on the final scheduling path and generates compensation factors based on customer feedback levels to correct logistics operation indicators during the execution process. When the corrected logistics operation indicators still show abnormalities after being compared with the historical indicator set again, emergency path reconstruction and resource reallocation are triggered in sequence. The early warning module generates tiered early warning information based on the revised logistics operation indicators. The tiered early warning information is divided into Level 1, Level 2 and Level 3 early warnings, and triggers corresponding risk warnings and handling procedures. The display module shows the logistics operation indicators, the candidate scheduling path set, the corrected logistics operation indicators, and the hierarchical early warning information.

Images