Intelligent logistics matching method based on dynamic data enabling

Abstract

The invention discloses an intelligent logistics matching method based on dynamic data enabling, and particularly relates to the technical field of logistics matching. The method comprises the following steps: firstly, acquiring real-time operation data of a logistics matching system, and constructing an initial state feature vector through standardization processing, secondly, screening out a target order delivery group which operates stably and is matched reasonably, and then, matching an optimal order delivery group by calculating the state feature similarity of a current order; aiming at a matching abnormal condition, adopting an abnormal mode identification method, extracting a transport capacity distribution abnormal characteristic and a time sequence change abnormal characteristic of a matching mode, calculating an abnormal matching degree, generating a matching adjustment instruction based on a calculation result, and dynamically optimizing a matching weight; the problem of deviation accumulation caused by self-enhancement of the matching algorithm is effectively avoided, the algorithm island phenomenon is reduced, the adaptive capacity of the logistics matching system is improved, logistics resource allocation is more efficient and fair, and the operation efficiency of the whole logistics network is improved.

Description

Technical Field

[0001] The present invention relates to the field of logistics matching technology, and in particular to an intelligent logistics matching method based on dynamic data empowerment. Background Art

[0002] With the rapid development of e-commerce, instant delivery and supply chain management, the demand for intelligent logistics matching is increasing. Traditional logistics matching methods mainly rely on static data, such as fixed transportation routes, preset warehouse resource allocation rules and historical order data. However, this method cannot fully cope with dynamic changes in the logistics environment, such as sudden traffic jams, real-time order demand fluctuations, weather conditions, etc. Therefore, intelligent logistics matching methods based on dynamic data empowerment have gradually become the focus of research and application. This method realizes accurate and efficient logistics matching by collecting dynamic data such as traffic flow, vehicle location, order status, etc. in real time, and combines artificial intelligence, optimization algorithms and big data analysis technology, thereby improving transportation efficiency, reducing costs and improving user experience.

[0003] The prior art has the following deficiencies: In an intelligent logistics matching system based on dynamic data, the algorithm will continuously adjust scheduling decisions to adapt to the real-time changing environment. However, if the system relies too much on the data generated by itself rather than the external real environment data, it may cause the problem of distortion in the dynamic data feedback loop. Specifically, the logistics matching algorithm may deviate due to continuous automatic adjustments in a short period of time. For example, due to the imbalance of early scheduling decisions, orders in certain areas are mistakenly judged as low-demand areas, thereby reducing the capacity allocation in the area, further exacerbating the imbalance between supply and demand. This cumulative error may be amplified after a long period of operation, resulting in long-term low logistics efficiency in some areas, and even the phenomenon of algorithm islands, that is, order matching in certain areas is ignored or treated with low priority for a long time, affecting the fairness and stability of the overall logistics network. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent logistics matching method based on dynamic data empowerment to solve the shortcomings of the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solution: an intelligent logistics matching method based on dynamic data empowerment, comprising: Obtain the real-time operation data of the logistics matching system and standardize it to form the initial state feature vector; Based on long-term monitoring data, select the target order delivery group with stable logistics network operation and reasonable matching results; Calculate the similarity between the initial state feature vector and the state feature of the target order delivery group, determine the best matching order delivery group, and generate the optimal logistics matching plan; If the similarity is lower than the preset similarity threshold, the logistics matching is determined to be abnormal, and different types of abnormal matching patterns are identified to extract abnormal features under each abnormal matching pattern; Combining the abnormal characteristics of logistics capacity distribution and the abnormal characteristics of the time series changes of the matching pattern, the abnormal matching degree is calculated, and the corresponding matching adjustment instructions are generated according to the calculation results to dynamically adjust the matching weight.

[0006] Preferably, the screening of the target order delivery group further adopts the K-means clustering algorithm to classify the historical order data to determine the target group with the most stable matching and the highest delivery success rate: set the number of clusters k, initialize the cluster center, and calculate the Euclidean distance from each order group to the cluster center: iteratively update the cluster center until convergence, and select the order group with the smallest standard deviation and the highest matching success rate as the target delivery group.

[0007] Preferably, the similarity between the initial state feature vector and the target order delivery group is calculated using a cosine similarity calculation method. If the cosine similarity is lower than a set threshold, the match is determined to be abnormal, and the abnormal matching pattern is identified, and abnormal characteristics of logistics capacity distribution and abnormal characteristics of matching pattern timing changes are extracted for abnormal pattern analysis and adjustment.

[0008] Preferably, after analyzing the abnormal changes in the extracted regional capacity supply-demand ratio, an abnormal capacity supply-demand ratio index is calculated. The calculation method of the abnormal capacity supply-demand ratio index is: In each time window t, the available capacity data of the target area A is collected , indicating the number of vehicles / riders currently available for dispatch in area A and the number of orders to be matched , indicating the total number of orders to be matched in area A; In each time window t, the regional capacity supply-demand ratio is calculated: ; Calculate the rate of change of the capacity supply-demand ratio within the time window , the expression is: ; Calculate the mean and standard deviation of the historical capacity supply-demand ratio of region A in the past N time windows, and define the capacity supply-demand ratio abnormality index, which is expressed as: ;in: is the abnormal index of transport capacity supply-demand ratio, To prevent extremely small positive numbers with a denominator of 0, is the historical average of the supply-demand ratio of transport capacity, is the standard deviation of the historical capacity supply-demand ratio.

[0009] Preferably, after analyzing the acquired abnormal matching mode conversion frequency, a matching mode conversion frequency deviation index is generated, and the matching mode conversion frequency deviation index is obtained by: Record the number of matching mode conversions in the past N time windows to form a time series, train the ARIMA model to enable it to predict the normal trend of the matching mode conversion frequency; calculate the actual matching mode conversion frequency and ARIMA forecasts The Difference , the expression is: ; After calculating the mean square error MSE of the difference, calculate the matching mode conversion frequency deviation index, the expression is: ; where c is a very small number, Frequency deviation index for matching mode conversion.

[0010] Preferably, the capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index are normalized so that they are both between [0,1], and the abnormal matching degree is calculated based on the normalized capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index.

[0011] Preferably, the obtained abnormal matching degree is compared with a preset threshold. If the abnormal matching degree is greater than or equal to the preset threshold, it means that there is an abnormality in the current matching strategy. At this time, a matching adjustment instruction is generated, including adjusting the matching weight, dynamically optimizing the capacity allocation, and optimizing the matching algorithm. If the abnormal matching degree is less than the preset threshold, it means that the current matching strategy is still within the normal range, and there is no need to adjust the matching plan, and the current matching rules are maintained.

[0012] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention constructs a standardized initial state feature vector, combines long-term monitoring data, and screens out target order delivery groups with stable matching and high delivery success rate to achieve accurate logistics matching. Through cosine similarity calculation, the current order is dynamically matched with the optimal order delivery group, and when the matching is abnormal, the capacity supply and demand ratio abnormality index and the matching mode conversion frequency deviation index are used for intelligent analysis to identify the imbalance of logistics capacity distribution and abnormal matching mode fluctuations, thereby optimizing the matching strategy. The present invention uses machine learning methods such as K-means clustering and ARIMA time series analysis to improve the stability of matching, avoid the accumulation of deviations caused by data feedback loop distortion, reduce the algorithm island phenomenon, and improve the fairness and overall operating efficiency of the logistics system.

[0013] 2. The present invention can monitor the status of the logistics network in real time, dynamically identify and optimize abnormal matching situations, and ensure that the matching strategy conforms to the actual supply and demand relationship. By normalizing the abnormal index and setting the threshold based on the abnormal matching degree, the matching strategy can be automatically adjusted and optimized, thereby improving the flexibility and adaptability of logistics scheduling. The present invention can optimize the allocation of logistics resources in a dynamic and complex environment, reduce the empty load rate, improve the distribution efficiency, reduce the cost, and ensure the long-term stable operation of the system, providing more accurate and reliable technical support for intelligent logistics matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0015] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] For examples, see Figure 1 As shown, the intelligent logistics matching method based on dynamic data empowerment described in this embodiment includes: Obtain the real-time operation data of the logistics matching system and standardize it to form the initial state feature vector; Based on long-term monitoring data, select the target order delivery group with stable logistics network operation and reasonable matching results; Calculate the similarity between the initial state feature vector and the state feature of the target order delivery group, determine the best matching order delivery group, and generate the optimal logistics matching plan; If the similarity is lower than the preset similarity threshold, the logistics matching is determined to be abnormal, and different types of abnormal matching patterns are identified to extract abnormal features under each abnormal matching pattern; Combining the abnormal characteristics of logistics capacity distribution and the abnormal characteristics of the time series changes of the matching pattern, the abnormal matching degree is calculated, and the corresponding matching adjustment instructions are generated according to the calculation results to dynamically adjust the matching weight.

[0018] The real-time operation data of the logistics network is obtained from various data sources of the logistics matching system (including order management system, Internet of Vehicles system, map API, weather forecast service, etc.). The collected data mainly includes the following categories: (1) Order information data: order number, order time, order source; order type (same-city delivery, long-distance transportation, express delivery, etc.); cargo attributes (weight, volume, special storage requirements); pickup location, delivery location, expected delivery time; (2) Transport capacity status data: real-time location and available load of transport vehicles; delivery routes and historical driving trajectories of vehicles; driver information (experience level, driving habits, etc.); empty load rate and current delivery task status; (3) Traffic condition data (obtained through map API): real-time road traffic conditions (unblocked, congested, closed, etc.); traffic density on major logistics arteries; accident reports, construction areas, and other influencing factors; (4) Environmental factor data: weather conditions (temperature, rainfall, wind speed, etc.); emergencies (natural disasters, road closure notices, etc.); the frequency of data collection needs to be set according to the real-time requirements of the logistics system. For example, high-frequency deliveries (such as food delivery, same-city express delivery) may need to be updated every minute, while long-distance transportation may need to be updated every hour.

[0019] The collected data often contains redundancy, errors and omissions, and needs to be preprocessed to improve data quality. Use fields such as order number and timestamp to remove duplicated order information. Optimize the trajectory of GPS location data and remove offset values. Use the Z-score method or box plot method to detect and remove outliers, such as: vehicle speed far exceeds the speed limit or is negative (abnormal GPS data); unify the format, such as converting the timestamp to a standard format (ISO 8601). Use mean imputation: For example, if the traffic flow data in a certain area is missing in a short period of time, it can be filled with the average value of the adjacent time period.

[0020] Different types of data may have different numerical ranges. For example, the number of orders may be from hundreds to thousands, while the traffic congestion index is between 0 and 1. In order to prevent the difference in numerical scales from affecting the matching algorithm, normalization is required. The normalized data needs to further extract key features to construct the feature vector required by the intelligent matching algorithm.

[0021] Convert the extracted features into vector form according to a certain format, for example: ;in: Represent order characteristics (e.g., order volume, delivery urgency); Represents capacity characteristics (such as load factor, route optimization degree); Represents environmental characteristics (such as weather impact, traffic index).

[0022] The final feature vector will be stored in the dynamic data cache of the matching system as the input of the intelligent matching algorithm. The vector will be continuously updated in the subsequent matching optimization steps to ensure that the matching results can adapt to the real-time changing logistics environment.

[0023] Extract long-term monitoring information from the historical data of the logistics matching system, mainly including: Order matching records: order quantity, delivery time, delivery success rate, historical matching strategy, etc. Capacity allocation: vehicle utilization rate, empty load rate, overload rate, delivery task completion, etc. Distribution network stability: delivery time fluctuation, order concentration, order distribution in the delivery area, etc. Arrange the data in a time series format to ensure that it contains enough historical records (such as data from the past 3 months, 6 months, and 1 year).

[0024] In order to screen out stable order delivery groups, the following key stability indicators need to be calculated: Calculate the proportion of historical orders that can be delivered on time after matching , the expression is: ;like >90% (threshold is adjustable), indicating a reasonable match and proceeding to the next step of screening.

[0025] Calculate the deviation between the actual delivery time and the estimated delivery time of the matched orders , measured by standard deviation, the expression is: ; Set thresholds (such as <15 minutes), if satisfied, the match is considered stable.

[0026] Calculate the distribution of orders in different regions If the number of orders is concentrated in a specific area (such as a city or a specific business district) for a long time, it means that the matching regularity is strong, and the Gini coefficient can be used to measure the order distribution: ;in, Represents the proportion of orders in different regions. If G<0.4, it means that the order distribution is relatively balanced and the matching is reasonable.

[0027] In order to further screen out the matching stable order groups, K-means clustering algorithm or DBSCAN density clustering algorithm can be used for data classification.

[0028] Construct the following feature vector: ;in: is the matching success rate, is the volatility of delivery time, G is the order density (distribution of delivery hotspots), is the average delivery distance, is the capacity load factor (empty / full load ratio), is the average traffic congestion index. Use the K-means algorithm to classify the order delivery groups and determine the stable matching groups: set the number of clusters k (such as 3-5), initialize the cluster center, calculate the Euclidean distance from each order group to the cluster center: iteratively update the cluster center until convergence. Select the order group with the smallest standard deviation and the highest matching success rate as the target delivery group.

[0029] In the intelligent logistics matching process, it is necessary to calculate the state feature vector of the current order based on dynamic data and perform similarity matching with the target order delivery group to determine the optimal matching solution.

[0030] Initial state eigenvector ;in: Represent order characteristics (e.g., order volume, delivery urgency); Represents capacity characteristics (such as load factor, route optimization degree); Represents environmental characteristics (such as weather impact, traffic index). This vector is processed by data standardization (such as Min-Max normalization) to make each feature value in the same scale.

[0031] The feature vector library S of the target order delivery group stores multiple historical matching stable group feature vectors: ; each A feature vector representing a group of order delivery orders with a high historical matching success rate.

[0032] Calculate the cosine similarity between the current order vector V and each target group vector S , the expression is: ; It is between [-1,1]. The closer the value is to 1, the more similar the current order is to the characteristics of the target group.

[0033] The target order delivery group is selected with the highest similarity. After the optimal matching order group is determined, the final logistics matching plan is generated, which includes: based on the historical data of the target order delivery group, the most matching transportation resources are selected: ensuring a reasonable capacity load rate to avoid overloading or empty load. Prioritize transportation plans with relatively stable historical delivery timeliness. If the target group involves multiple delivery areas, cluster scheduling optimization is used to divide the orders into reasonable transportation batches.

[0034] Based on the historical trajectory data of the target delivery group and combined with real-time traffic information, the A* search algorithm or Dijkstra algorithm is used to optimize the delivery path: if the current road is congested, an alternative path used by similar orders in history is selected. Combined with the order timeliness requirements, the delivery order is dynamically adjusted.

[0035] In special cases, the system dynamically adjusts the matching plan: if the similarity of the matching groups is insufficient, human-machine collaborative intervention is allowed to manually adjust the matching weight. Combined with historical trend analysis, the matching parameters are adjusted to improve matching accuracy. During peak hours, the capacity allocation is dynamically adjusted to ensure the best matching results.

[0036] When the calculated similarity is lower than the preset similarity threshold (for example, 0.8), it indicates that the matching scheme of the current order may be abnormal, and it is necessary to further identify abnormal matching patterns and extract key abnormal features under different abnormal patterns. Abnormal matching patterns can be divided into abnormal logistics capacity distribution and abnormal timing changes of matching patterns.

[0037] Abnormal matching patterns are automatically identified mainly through cluster analysis, statistical analysis, machine learning and other methods. The specific steps are as follows: K-means clustering to identify abnormal patterns: Use historical matching data as a training set and use the K-means clustering algorithm to classify matching patterns. Set the center cluster of the normal matching pattern. If the matching scheme of the current order deviates far from the center cluster, it is considered abnormal. Calculate the Euclidean distance between the current order matching scheme and each matching pattern. If it exceeds the standard deviation range of the normal pattern, it is classified as an abnormal match.

[0038] Anomaly detection based on statistical thresholds: Calculate the matching success rate: If it is lower than the historical average (such as 90%), it is considered abnormal. Calculate the average delivery time deviation: If it exceeds 3 times the standard deviation, it is an abnormal matching mode. Calculate the growth rate of failed matching orders. If the number of failed matching orders surges in a short period of time, it indicates that there is a problem with the system matching logic.

[0039] Anomaly detection based on machine learning (LOF method): The local anomaly factor algorithm is used to calculate the local anomaly factor score of the current matching scheme in the historical data. If the LOF score is high (much higher than normal orders), it indicates that the matching pattern is abnormal.

[0040] When the supply and demand of logistics capacity do not match, it may lead to a decrease in distribution efficiency. Extract key abnormal features: regional capacity supply and demand ratio , the expression is: ; If the supply-demand ratio of a region's transport capacity is far lower than the historical average, it means that there is a shortage of transport capacity in that region.

[0041] After analyzing the abnormal changes in the extracted regional capacity supply-demand ratio, the capacity supply-demand ratio abnormality index is calculated. The calculation method of the capacity supply-demand ratio abnormality index is: In each time window (e.g., 10 minutes, 30 minutes, 1 hour), collect the following data for target area A: Available capacity , indicating the number of vehicles / riders currently available for dispatch in area A. Number of orders to be matched , indicating the total number of orders currently to be matched in area A.

[0042] In each time window t, the regional capacity supply-demand ratio is calculated: ;like >1, indicating sufficient capacity, and there may be idle capacity. <1, indicating insufficient transport capacity, which may result in delayed order delivery.

[0043] Calculate the rate of change of the capacity supply-demand ratio within the time window , the expression is: ;like If the fluctuation is too large, it means that the regional supply and demand are changing dramatically and there may be abnormalities in the matching system.

[0044] Calculate the mean and standard deviation of the historical capacity supply-demand ratio of region A in the past N time windows, and define the capacity supply-demand ratio abnormality index to measure the abnormality of the current supply-demand ratio. The expression is: ;in: It is the abnormal index of capacity supply-demand ratio. The larger the value, the more serious the imbalance between supply and demand. To prevent extremely small positive numbers with a denominator of 0 (such as 0.0001). is the historical average of the supply-demand ratio of transport capacity, It is the standard deviation of the historical capacity supply-demand ratio, which measures the normal fluctuation range of supply and demand changes.

[0045] like <1, the supply-demand ratio is within the normal range and no adjustment is required. )<2, the supply-demand ratio is moderately abnormal, and the capacity scheduling can be adjusted appropriately. ≥2, the supply-demand ratio is abnormally serious, and the matching strategy needs to be adjusted immediately (such as increasing capacity and optimizing matching rules).

[0046] The matching logic of the logistics matching system should remain relatively stable. If the matching pattern fluctuates abnormally in a short period of time, it indicates that there may be problems with the matching strategy. Extract key abnormal features: matching pattern conversion frequency , the expression is: ; If the matching mode switches frequently in a short period of time (such as switching 10 times within 5 minutes), it indicates that the system has a strategy instability problem.

[0047] After analyzing the abnormal situation of the acquired matching mode conversion frequency, a matching mode conversion frequency deviation index is generated. The matching mode conversion frequency deviation index is obtained by: Record the number of matching mode conversions in the past N time windows (such as 5 minutes, 10 minutes, 30 minutes) to form a time series, and select appropriate parameters (p, d, q), where: p is the autoregressive term, which indicates how many past values ​​the current value is related to. d is the difference order, which ensures the stability of the time series. q is the moving average term, which indicates the impact of the error on the current value. Train the ARIMA model to enable it to predict the normal trend of the matching mode conversion frequency.

[0048] Calculate the actual matching mode switching frequency and ARIMA forecasts The Difference , the expression is: ; After calculating the mean square error MSE of the difference, calculate the matching mode conversion frequency deviation index, the expression is: ; where c is a very small number (such as 0.0001) to prevent the denominator from being 0. is the matching mode conversion frequency deviation index; if ≥2, it indicates that the matching mode conversion frequency is seriously abnormal and the matching strategy may need to be re-optimized.

[0049] The capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index are normalized so that they are both between [0,1]. The abnormal matching degree is calculated based on the normalized capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index.

[0050] For example, the present invention can use the following formula to calculate the abnormal matching degree, and the calculation expression is: ; In the formula, is the abnormal matching degree, is the abnormal index of transport capacity supply-demand ratio, is the matching mode conversion frequency deviation index, is the weight coefficient of the abnormal index of the capacity supply-demand ratio and the matching mode conversion frequency deviation index (which can be optimized based on experimental experience or machine learning), and Both are greater than 0.

[0051] Compare the obtained abnormal matching degree with the preset threshold. If the abnormal matching degree is greater than or equal to the preset threshold, it means that the current matching strategy is abnormal, which may be caused by uneven capacity distribution or unstable matching mode. Generate matching adjustment instructions: Adjust matching weight: If If the capacity is large (imbalance in transport capacity), increase the weight of transport capacity scheduling and optimize the supply and demand matching strategy. Larger (matching patterns fluctuate frequently), reducing the frequency of algorithm adjustments and enhancing the stability of matching strategies.

[0052] Dynamic optimization of capacity allocation: If the supply-demand ratio is seriously unbalanced, more capacity will be dispatched to the shortage area. If there is a long-term imbalance between supply and demand in a certain area, the long-term matching rules will be adjusted (such as optimizing storage locations and adding transfer points).

[0053] Optimize matching algorithms: Combine historical data to retrain matching models to improve matching stability and accuracy. Reduce the sensitivity of high-frequency strategy switching and reduce unnecessary mode conversions.

[0054] If the abnormal matching degree is less than the preset threshold, it means that the current matching strategy is still within the normal range, and there is no need to adjust the matching plan, and the current matching rules are maintained.

[0055] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0056] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0057] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the objects associated with each other are in an "or" relationship, but it may also indicate an "and / or" relationship, which can be understood by referring to the context. A person of ordinary skill in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this article can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0058] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. Intelligent logistics matching method based on dynamic data empowerment, characterized by: include: Obtain the real-time operation data of the logistics matching system and standardize it to form the initial state feature vector; Based on long-term monitoring data, select the target order delivery group with stable logistics network operation and reasonable matching results; Calculate the similarity between the initial state feature vector and the state feature of the target order delivery group, determine the best matching order delivery group, and generate the optimal logistics matching plan; If the similarity is lower than the preset similarity threshold, the logistics matching is determined to be abnormal, and different types of abnormal matching patterns are identified to extract abnormal features under each abnormal matching pattern; Combining the abnormal characteristics of logistics capacity distribution and the abnormal characteristics of the time series changes of the matching pattern, the abnormal matching degree is calculated, and the corresponding matching adjustment instructions are generated according to the calculation results to dynamically adjust the matching weight.

2. The intelligent logistics matching method based on dynamic data empowerment according to claim 1 is characterized by: The screening of the target order delivery group further adopts the K-means clustering algorithm to classify the historical order data to determine the target group with the most stable matching and the highest delivery success rate: set the number of clusters k, initialize the cluster center, and calculate the Euclidean distance from each order group to the cluster center: iteratively update the cluster center until convergence, and select the order group with the smallest standard deviation and the highest matching success rate as the target delivery group.

3. The intelligent logistics matching method based on dynamic data empowerment according to claim 2 is characterized by: The similarity between the initial state feature vector and the target order delivery group is calculated using a cosine similarity calculation method. If the cosine similarity is lower than a set threshold, the match is determined to be abnormal, and the abnormal matching pattern is identified. The abnormal characteristics of the logistics capacity distribution and the abnormal characteristics of the matching pattern time series changes are extracted for abnormal pattern analysis and adjustment.

4. The intelligent logistics matching method based on dynamic data empowerment according to claim 3 is characterized by: After analyzing the abnormal changes in the extracted regional capacity supply-demand ratio, the capacity supply-demand ratio abnormality index is calculated. The calculation method of the capacity supply-demand ratio abnormality index is: In each time window t, the available capacity data of the target area A is collected , indicating the number of vehicles / riders currently available for dispatch in area A and the number of orders to be matched , indicating the total number of orders to be matched in area A; In each time window t, the regional capacity supply-demand ratio is calculated: ; Calculate the rate of change of the capacity supply-demand ratio within the time window , the expression is: ; Calculate the mean and standard deviation of the historical capacity supply-demand ratio of region A in the past N time windows, and define the capacity supply-demand ratio abnormality index, which is expressed as: ;in: is the abnormal index of transport capacity supply-demand ratio, To prevent extremely small positive numbers with a denominator of 0, is the historical average of the supply-demand ratio of transport capacity, is the standard deviation of the historical capacity supply-demand ratio.

5. The intelligent logistics matching method based on dynamic data empowerment according to claim 4 is characterized in that: After analyzing the abnormal situation of the acquired matching mode conversion frequency, a matching mode conversion frequency deviation index is generated. The matching mode conversion frequency deviation index is obtained by: Record the number of matching mode conversions in the past N time windows to form a time series, train the ARIMA model to enable it to predict the normal trend of the matching mode conversion frequency; calculate the actual matching mode conversion frequency and ARIMA forecasts The Difference , the expression is: ; After calculating the mean square error MSE of the difference, the matching mode conversion frequency deviation index is calculated, and the expression is: ; where c is a very small number, Frequency deviation index for matching mode conversion.

6. The intelligent logistics matching method based on dynamic data empowerment according to claim 5 is characterized by: The capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index are normalized so that they are both between [0,1]. The abnormal matching degree is calculated based on the normalized capacity supply-demand ratio anomaly index and the matching mode conversion frequency deviation index.

7. The intelligent logistics matching method based on dynamic data empowerment according to claim 6 is characterized by: The obtained abnormal matching degree is compared with the preset threshold. If the abnormal matching degree is greater than or equal to the preset threshold, it means that there is an abnormality in the current matching strategy. At this time, a matching adjustment instruction is generated, including adjusting the matching weight, dynamically optimizing the capacity allocation, and optimizing the matching algorithm. If the abnormal matching degree is less than the preset threshold, it means that the current matching strategy is still within the normal range, and there is no need to adjust the matching plan, and the current matching rules are maintained.

Images