Logistics transportation management method and system based on multi-source data fusion

Abstract

The invention relates to the technical field of logistics management, in particular to a logistics transportation management method and system based on multi-source data fusion, and the method comprises the following steps: monitoring the vehicle transportation duration in real time, calculating a congestion point, judging a coincidence time period in combination with predicted replenishment time, and generating a rhythm window; and matching the storage delivery time and sorting to generate a path segment code, fitting a path to screen a vehicle binding label, sending an instruction to monitor the path consistency, and generating a scheduling result. According to the invention, transportation time is monitored in real time, congestion points are generated, a rhythm window is dynamically generated in combination with replenishment task time, path segment coding is carried out on a picking sequence, channel use and shipping time, vehicle resources are screened based on path fitting and a path label is bound, and distribution consistency is monitored in real time to generate a scheduling instruction. Multi-dimensional data are analyzed in a unified mode and driven in a time sequence mode, intervention is reduced, efficiency is improved, path coding matching tracks are combined with congestion adjustment to avoid conflicts and redundancy, and the response period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics management, and in particular to a logistics transportation management method and system based on multi-source data fusion. Background Art

[0002] The field of logistics management encompasses technologies related to planning, organizing, coordinating, and controlling the entire process of goods delivery, from supply to demand. Its core focus is improving the operational efficiency and responsiveness of logistics systems through scientific scheduling and resource allocation, encompassing transportation, warehousing, loading, unloading, and distribution. This technology, grounded in information management, integrates key elements such as route optimization, inventory management, and operational process control to form a comprehensive management system encompassing procurement planning, transportation execution, order management, and customer service. This system is widely used across various industries, including manufacturing, e-commerce, and retail, and is continuously expanding into smart logistics and collaborative supply chain management.

[0003] Among them, the logistics and transportation management method based on multi-source data fusion refers to the realization of information coordination and management of the entire logistics and transportation task process by unified integration of information from the transportation vehicle positioning system, warehousing equipment, logistics order platform and external environment data sources. It mainly aims at the problems of data fragmentation, update delay and response lag in traditional logistics and transportation. It adopts data time synchronization technology, information entity association technology and path status fusion calculation means to unify modeling and real-time fusion of logistics transportation path, transportation carrier delivery time and operation status data. Its participation content includes the standardization of data format between logistics nodes, matching verification between transportation vehicle operation trajectory and planned tasks, multi-dimensional time sequence arrangement of logistics information and rule-driven transportation process reconstruction, which belongs to the information fusion and path planning collaboration subcategory in intelligent logistics decision support system.

[0004] Traditional logistics management relies on static route planning and discrete data sources. Links like vehicle location, warehouse timing, and order status lack real-time data interoperability, resulting in inefficient cross-node collaboration. For example, when a vehicle needs to reroute due to sudden congestion, the system cannot quickly correlate warehouse operation status with available capacity, making manual coordination time-consuming and error-prone. Existing technologies use fixed time windows to schedule replenishment tasks, failing to consider the impact of dynamic changes in transportation routes on warehouse resources. For example, conflicts between replenishment windows and picking lanes can easily cause warehouse operations to stall. Verification of vehicle route records and transportation tasks relies on offline data, lacking real-time route encoding for fitting and analysis, resulting in high resource allocation redundancy. For example, multiple empty vehicles are repeatedly dispatched in the same area, or capacity is insufficient on high-demand routes. These issues arise from a lack of established temporal correlation rules between transportation tasks and warehouse operations. The data processing lags inherent in existing technologies lead to a disconnect between transportation and warehousing, making it difficult to support the flexible scheduling requirements of complex scenarios. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the embodiment of the present invention provides a logistics and transportation management method and system based on multi-source data fusion. The technical solution includes the following:

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: a logistics and transportation management method based on multi-source data fusion, comprising the following steps:

[0007] S1: Real-time monitoring of the transport time of logistics vehicles, inputting the transport time of the vehicles into the principal component analysis model, and calculating and recording the congestion score of the transport route;

[0008] S2: Obtain the estimated arrival start and end times of the replenishment tasks within the next transportation cycle, determine the overlapping replenishment time periods based on the congestion score of the transportation route, perform shifting based on the start and end times to generate multiple rhythm window candidate time periods, and determine the replenishment rhythm window;

[0009] S3: Matching the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse-side scheduled shipping time, sorting the warehouse-side picking and delivery sequence, transportation channel usage sequence, and scheduled shipping time data using a dynamic priority scheduling algorithm to generate a path segment sequence code;

[0010] S4: Calling the path segment sequence code to calculate the fitting degree of the vehicle path sequence record collected by the vehicle GPS, screening vehicle resources of the same sequence and binding the corresponding vehicle delivery path label;

[0011] S5: Based on the replenishment rhythm window matching judgment, a delivery instruction is sent, the consistency between the delivery path and the vehicle delivery path label is monitored in real time, and a transportation scheduling adjustment result is generated according to the monitoring result.

[0012] As a further solution of the present invention, the node congestion score includes vehicle transportation time, route congestion index, and node load compensation value; the replenishment rhythm window includes the rhythm window start time, the rhythm window end time, and the rhythm window alternative time period; the path segment sequence code is specifically the picking sort code, the channel occupancy sequence number, and the shipping time sequence number; the vehicle delivery path label includes the path sequence fitting degree, the vehicle matching sequence number, and the path binding identifier; the transportation scheduling adjustment result includes the distribution path matching degree, the vehicle path consistency, and the real-time scheduling deviation.

[0013] As a further embodiment of the present invention, the specific steps of monitoring the transport time of logistics vehicles in real time, inputting the vehicle transport time into a principal component analysis model, calculating the congestion score of the transport route, supplementing the node load score data, and generating the congestion score include the following:

[0014] S101: monitoring the start and end times of a transport task of a logistics vehicle, and analyzing the duration of a single vehicle operation of the transport task based on the deviation between the start and end times of the transport task obtained by monitoring, to obtain the transport duration;

[0015] S102: Based on the transport duration input into the principal component analysis model, deviation comparison is performed in combination with a preset duration threshold range of the corresponding transport route to establish a route congestion analysis result;

[0016] As a further solution of the present invention, it is characterized in that a deviation comparison is performed on a preset time threshold range of the transport route, using the formula:

[0017]

[0018] Calculate the congestion level I of the node's current route and obtain the congestion level analysis result of the node's current route;

[0019] Where n is the total number of sample vehicles, i is the index of the vehicle, and U T =μ T +2σ T is the upper threshold, L T =μ T -2σ T is the lower threshold, μ T is the average transport time of transport vehicles, μ is the standard sign of the mean, σ is the standard deviation, δ is the penalty coefficient, ε is the reward coefficient, μ+2σ is the upper threshold, μ-2σ is the lower threshold, max(T i -U T ,0) means that if the transportation time exceeds the upper threshold, the corresponding reward is calculated, max(L T -T i ,0) means if the transport time is lower than the lower threshold, the corresponding penalty is calculated, C adj represents the average congestion index of neighboring nodes, C base Expressed as the baseline congestion index.

[0020] S103: performing point mapping processing on the nodes associated with the transport route according to the route congestion analysis result, and generating a congestion point of the node's current transport route.

[0021] As a further solution of the present invention, the estimated replenishment arrival start and end times of the replenishment task within the next transportation cycle are obtained, and the overlapping replenishment time periods are determined in combination with the congestion score of the transportation route. Based on the start and end times, a shift is performed to generate multiple candidate rhythm window time periods. The specific steps of determining the replenishment rhythm window include:

[0022] S201: Arrival Time: Obtain the estimated arrival time and departure time corresponding to each replenishment task in the next transportation cycle, classify the start and end times of each replenishment task according to the transportation cycle number, and locate the time boundaries of the start and end time intervals of the replenishment tasks in each cycle to generate the replenishment time interval;

[0023] S202: Based on the obtained replenishment time interval, the node congestion score is called to determine whether each replenishment time interval coincides with the score of the corresponding node, and replenishment tasks with overlapping time periods are filtered and aggregated to obtain conflicting time intervals.

[0024] S203: Based on the conflicting time interval, the replenishment time interval is called, and time axis shift processing is performed on the boundaries of both ends of the conflicting time interval, including adjusting the starting boundary time and the ending boundary time of the conflicting time interval as a shift operation, to generate multiple non-overlapping new time intervals, and eliminating windows that intersect with the start and end times of the replenishment task to obtain rhythm candidate intervals;

[0025] S204: Based on the obtained rhythm candidate intervals, all candidate intervals are arranged in ascending order according to the start time, and time window combinations that do not overlap within a single cycle are screened to generate a replenishment rhythm window.

[0026] As a further solution of the present invention, the replenishment estimated arrival start time and the replenishment estimated arrival end time in the replenishment rhythm window are matched and judged with the warehouse-side scheduled shipping time, and the warehouse-side picking and delivery sequence, the transportation channel usage sequence, and the scheduled shipping time data are serialized and sorted. The specific steps of generating the path segment sequence code include:

[0027] S301: Obtain the replenishment estimated arrival start time and the replenishment estimated arrival end time in the replenishment rhythm window, match the start and end times of the replenishment tasks in the replenishment rhythm window with the scheduled shipping time of the warehouse end, call the warehouse task node information to perform logical judgment on the matching group, group the task combinations with overlapping time into the same rhythm group, and generate a time matching result between the replenishment tasks and the warehouse nodes;

[0028] S302: Based on the obtained time matching result between the replenishment task and the warehouse node, the warehouse picking plan number, transportation channel code, and shipping appointment time corresponding to the task in the rhythm group are called, and the three types of data are bound to perform task association sequence. The parameters in the task group are jointly sorted according to the time sequence relationship to obtain the operation rhythm sorting sequence of the associated nodes after sorting;

[0029] S303: According to the generated operation rhythm sorting sequence, the warehouse number, channel number and shipping number in the sorting sequence are called, and the path segment structure information is generated in sequence. The path fields in the same rhythm group are marked with unique picking sorting codes, and the sequence segment number is established and the corresponding task path mapping relationship is recorded to generate the path segment sequence code.

[0030] As a further solution of the present invention, the specific steps of calling the path segment sequence code to perform similarity calculation on the vehicle path sequence record collected by the vehicle GPS, screening vehicle resources with similar sequences and binding corresponding vehicle path labels include:

[0031] S401: Calling the path segment sequence code, obtaining the vehicle path sequence record collected by the vehicle-mounted GPS, comparing the matching degree of the three types of fields in the path segment, namely, the numbered node, the time sequence, and the channel number, calculating the structural coincidence rate between the current path segment and the preset reference value of the vehicle path segment, and generating the path fit;

[0032] As a further solution of the present invention, it is characterized in that the preset reference values ​​of the current path segment and the vehicle path segment are processed using the formula:

[0033]

[0034] Calculate the structural overlap rate S overlap ;

[0035] Among them, N match It is the node number set corresponding to the real-time data of the vehicle's current path status and the preset benchmark of the vehicle path segment. The number of node number matches formed by counting the number of identical numbers is T. max is the time difference tolerance threshold, which is obtained by statistical quantiles, N total Indicates the total number of nodes in the current path segment. represents the mean of the time sequence difference, C match Indicates the number of channel number matches, C total Indicates the total number of channels in the current path segment.

[0036] S402: Based on the acquired path fit, the idle time and available loading status parameters of each vehicle in the vehicle status list in the current cycle are retrieved, a path matching threshold is determined by the path overlap rate, and vehicles with path fits higher than the path matching threshold and meeting the status conditions are selected as a candidate set to establish vehicle matching resources;

[0037] S403: Based on the obtained vehicle matching resources, the path matching threshold is determined by calculating the distribution characteristics through the time difference data, and the unique identification code of each candidate vehicle and the corresponding path segment sequence code are called to generate a label combination. The combined labels are grouped by vehicle and written into the vehicle task allocation field to establish a vehicle deployment path label.

[0038] As a further solution of the present invention, based on the replenishment rhythm window matching judgment, sending a delivery instruction, monitoring the consistency of the delivery path and the vehicle placement path label in real time, and generating a transportation scheduling adjustment result based on the monitoring result include the following specific steps:

[0039] S501: Obtain the arrival start and end times of the replenishment task in the replenishment rhythm window, call the vehicle delivery path tag, match the start and end times of the replenishment task with the path segment times recorded in the corresponding vehicle path tag, and measure the time difference to generate a rhythm matching offset result for the current path task;

[0040] S502: Based on the obtained rhythm matching offset result, the route task numbers with time offsets less than the offset threshold are screened, the corresponding vehicle numbers and replenishment task numbers are called to execute the combinations, a delivery instruction is generated for each combination, the route number and issuance time are recorded, and a delivery task is established;

[0041] S503: Calling the obtained delivery task, collecting the current positioning trajectory of the delivery vehicle and synchronously comparing it with the preset path segment sequence number, extracting the current vehicle operation number of each path segment and the delivery tag path segment number to perform node consistency judgment, obtaining the ratio of the number of consistent nodes in each path segment to the total number, and generating a path consistency ratio;

[0042] S504: Based on the obtained path consistency ratio, the path segment sequence code is called, the path segment number is replaced based on the real-time scheduling deviation of the path deviation benchmark ratio, the path segment offset result is replaced and recorded and appended to the control scheduling log to generate the transportation scheduling adjustment result.

[0043] On the other hand, a logistics and transportation management system based on multi-source data fusion is provided, which is applied to a logistics and transportation management method based on multi-source data fusion, and the system includes:

[0044] The node load integration module monitors the transport time of logistics vehicles in real time, inputs the vehicle transport time into the principal component analysis model, calculates the congestion score of the transport route, supplements the node load integration data, and transmits the generated congestion score to the replenishment rhythm window module;

[0045] The replenishment rhythm window module obtains the estimated replenishment arrival start and end times of the replenishment task in the next transportation cycle, determines the overlapping replenishment time periods based on the congestion score of the transportation route, performs shifting based on the start and end times to generate multiple rhythm window candidate time periods, and passes them to the path segment sequence generation module;

[0046] The path segment sequence generation module matches the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse-side shipping scheduled time, and serializes and sorts the warehouse-side picking and outbound order, the transportation channel usage order, and the shipping scheduled time data to generate a path segment sequence code and pass it to the vehicle label binding module;

[0047] The vehicle label binding module calls the current path segment sequence in the path segment sequence code, compares the fit with the path segment sequence of the previous period, selects vehicle resources with similar sequences, binds the corresponding vehicle delivery path labels, and transmits them to the collaborative scheduling control module;

[0048] The collaborative scheduling control module sends delivery instructions based on the replenishment rhythm window matching judgment, monitors the consistency of the delivery path and the vehicle delivery path label in real time, and generates transportation scheduling adjustment results based on the monitoring results.

[0049] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0050] Real-time monitoring of transport times is used to generate congestion scores based on principal component analysis models. Combined with the start and end times of replenishment tasks, a dynamic rhythm window is generated to precisely align transport demand with warehousing operations. Path segment sequence coding is used to time-sequence the picking sequence, handling channel usage, and shipping time, ensuring the coordinated matching of warehousing resources and transport tasks. Resources are selected based on the fit calculation of vehicle path sequence records. After binding path labels, delivery consistency is monitored in real time, allowing for rapid generation of scheduling adjustment instructions. This method integrates multi-dimensional data such as transport time, congestion status, and warehousing operation processes into a unified analysis framework. Through time-sequence coding and dynamic rule-driven approaches, the need for manual intervention is reduced. The coordinated optimization of transport routes and warehousing operations reduces vehicle waiting times and warehouse congestion, improving resource utilization. Similarity matching technology between path segment codes and vehicle trajectories, combined with a dynamic congestion score adjustment mechanism, proactively avoids route conflicts and capacity redundancy, shortening delivery response times. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0053] Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0055] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0056] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction between them is not emphasized, their intended meanings are the same.

[0057] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0058] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0059] See also Figure 1 The present invention provides a technical solution, a logistics and transportation management method based on multi-source data fusion, comprising the following steps:

[0060] S1: Real-time monitoring of the transportation time of logistics vehicles, inputting the vehicle transportation time into the principal component analysis model, and calculating and recording the congestion score of the transportation route;

[0061] S2: Obtain the estimated arrival start and end times of replenishment tasks within the next transportation cycle, determine the overlapping replenishment time periods based on the congestion score of the transportation route, perform displacement based on the start and end times, generate multiple rhythm window candidate time periods, and determine the replenishment rhythm window;

[0062] S3: Match the replenishment estimated arrival start and end times in the replenishment rhythm window with the warehouse-side shipping schedule. Use a dynamic priority scheduling algorithm to sort the warehouse-side picking and delivery sequence, the transportation channel usage sequence, and the shipping schedule data to generate a path segment sequence code.

[0063] S4: Call the path segment sequence code to calculate the fitting degree of the vehicle path sequence record collected by the vehicle GPS, select vehicle resources of the same sequence and bind the corresponding vehicle delivery path label;

[0064] S5: Based on the replenishment rhythm window matching judgment, the delivery instruction is sent, the consistency between the delivery path and the vehicle delivery path label is monitored in real time, and the transportation scheduling adjustment result is generated according to the monitoring result.

[0065] The node congestion score includes the vehicle transportation time, route congestion index, and node load compensation value. The replenishment rhythm window includes the rhythm window start time, rhythm window end time, and rhythm window alternative time period. The path segment sequence code is specifically the picking sort code, channel occupancy sequence number, and shipping time sequence number. The vehicle deployment path label includes the path sequence fitting degree, vehicle matching sequence number, and path binding identifier. The transportation scheduling adjustment results include the distribution path matching degree, vehicle path consistency, and real-time scheduling deviation.

[0066] See also Figure 1 ,real-time monitoring of the transportation time of logistics vehicles, inputting the vehicle transportation time into the principal component analysis model, calculating the congestion score of the transportation route, supplementing the node load score data, and generating the congestion score. The specific steps include:

[0067] S101: monitoring the start and end times of a transport task of a logistics vehicle, and analyzing the duration of a single vehicle operation of the transport task based on the deviation between the start and end times of the transport task obtained by monitoring, to obtain the transport duration;

[0068] Call the vehicle GPS module to collect the transportation task start time t of vehicle V001 in real time start =08:15:00 and end time t end =11:45:00, convert the timestamp to minutes t start =8×60+15=495 minutes, t end =11×60+45=705 minutes, calculate the time difference Δt=705-495=210 minutes, and check whether the timestamp synchronization error exceeds the threshold τ sync= 5 seconds. If the clock deviation between GPS modules is 3 seconds, the data is retained. If the deviation is 7 seconds, it is marked as abnormal data and needs to be re-collected. The same operation is performed on 5 sample vehicles to obtain the duration data set {210, 195, 225, 180, 204}, excluding those that exceed the physical limit value T max =480 minutes of data, and output the effective transportation duration result.

[0069] Table 1 Transport time monitoring data

[0070]

[0071]

[0072] S102: Based on the transport duration input into the principal component analysis model, deviation comparison is performed in combination with the preset duration threshold range of the corresponding transport route to establish a route congestion analysis result;

[0073] Call the transport duration dataset {210, 195, 225, 180, 204} for route R001 and calculate the mean Minutes, standard deviation Minutes, set the upper threshold U T =μ T +2σ T =202.8+2×16.5=235.8 minutes, lower threshold L T =μ T -1.5σ T = 202.8-1.5×16.5=178.05 minutes, loading the number of matching road segments N of the adjacent route R002 match =12, total number of road sections N total =15, time difference Δt=225-202.8=22.2 minutes, maximum tolerance T max = 40 minutes (based on the 95% percentile of historical data), matching congestion point C match =8, total congestion points C total =10, substitute into the formula to calculate the coverage score:

[0074]

[0075] Comparison coverage threshold S base =0.6. Since 0.682>0.6, it is determined that routes R001 and R002 have significant spatiotemporal correlation, and the congestion analysis result database is updated.

[0076] Table 2 Route matching parameters

[0077]

[0078]

[0079] Calculate the congestion level I of the node's current route and obtain the congestion level analysis result of the node's current route;

[0080] Where n is the total number of sample vehicles, i is the index of the vehicle, and U T =μ T +2σ T is the upper threshold, L T =μ T -2σ T is the lower threshold, μ T is the average transport time of transport vehicles, μ is the standard sign of the mean, σ is the standard deviation, δ is the penalty coefficient, ε is the reward coefficient, μ+2σ is the upper threshold, μ-2σ is the lower threshold, max(T i -U T ,0) means that if the transportation time exceeds the upper threshold, the corresponding reward is calculated, max(L T -T i ,0) means if the transport time is lower than the lower threshold, the corresponding penalty is calculated, C adj represents the average congestion index of neighboring nodes, C base Expressed as the baseline congestion index.

[0081] S overlap ≥0.6: Determined as a strong correlation, triggering cross-route congestion warning and path collaborative optimization. overlap <0.6: Determined as a weak association, only local path adjustment is performed.

[0082] The results show that the coverage score of route R001 is S overlap =0.682 exceeds the benchmark threshold, and the real-time congestion data of R002 (such as a 20% decrease in average vehicle speed) needs to be synchronized to R001.

[0083] S103: Based on the route congestion analysis results, perform point mapping processing on the nodes associated with the transport route to generate a congestion score for the node's current transport route;

[0084] Based on the traffic monitoring platform, the average speed v of the vehicles associated with the node is obtained in real time avg and delay time t delay , call the predefined congestion level judgment rules in the data, and set v avg With t delay Respectively with the preset threshold v low =40km / h, v mid =20km / h, t low = 5 minutes, t mid = 15 minutes for comparison, if v avg≥40 and t delay ≤5, it is judged as low congestion level. If v avg In the interval [20,40) or t delay Located in the interval (5,15], it is judged as medium congestion level. If v avg <20 or t delay >15, it is judged as high congestion level, and points are allocated for each level low =1,s mid =3,s high =5, call the road type database associated with the node and extract the road type parameter w i (Main Road i =1.5, secondary road w i =1.0, branch w i =0.5), calculate the node integral S i =s i ×w i , traverse all nodes in the route, and S i Accumulated as total score S total =∑S i For example, the real-time data of node N004 is v avg =35km / h, t delay = 8 minutes, because v avg ∈[20,40) and t delay ∈(5,15], judged as medium congestion level, s i =3, the road type is secondary road, w i =1.0, calculate S i =3×1.0=3.0, the real-time data of node N005 is v avg =18km / h, t delay = 20 minutes, because v avg <20 and t delay >15, it is judged as high congestion level, s i =5, road type is main road, w i =1.5, calculate S i =5×1.5=7.5, add to the total score;

[0085] Table 3 Example of node integral accumulation table

[0086] Node ID Congestion level (score sisi) Road type (weight wiwi) Points SiSi N004 Chinese (3) Secondary road (1.0) 3.0 N005 High (5) Main road (1.5) 7.5 N006 Low (1) Branch (0.5) 0.5

[0087] As shown in Table 3, the total score S total =3.0+7.5+0.5=11.0, weight coefficient w iThe setting basis is: the average impact factor of the main road in 100 congestion events is 2.4 minutes / km, the secondary road is 1.6 minutes / km, and the branch road is 0.8 minutes / km. The impact factor is normalized to the main road w i =2.4 / 1.6=1.5, secondary trunk road w i =1.6 / 1.6=1.0, branch w i =0.8 / 1.6=0.5, threshold S threshold = 10.0 is set by the score distribution of high-congestion routes in the statistical data. For example, 50 high-congestion routes are selected, and their total score mean is 12.3 and the standard deviation is 2.1. The threshold is set as mean minus standard deviation 12.3-2.1=10.2, rounded to 10.0. If S total ≥10.0, triggers the path optimization instruction.

[0088] See also Figure 1 The estimated arrival start and end times of replenishment tasks within the next transportation cycle are obtained. The overlapping replenishment time periods are determined based on the congestion score of the transportation route. Multiple candidate rhythm window periods are generated based on the shifting of the start and end times. The specific steps for determining the replenishment rhythm window include:

[0089] S201: Arrival Time: Obtain the estimated arrival time and departure time corresponding to each replenishment task in the next transportation cycle, classify the start and end times of each replenishment task according to the transportation cycle number, and locate the time boundaries of the start and end time intervals of the replenishment tasks in each cycle to generate the replenishment time interval;

[0090] Based on the task scheduling database of the transportation management system, the estimated arrival time T of all replenishment tasks in the next transportation cycle is extracted. arrive and the estimated departure time T depart , call the transport cycle division rule (every 4 hours is a cycle number C j , the cycle start time is 00:00, 04:00, 08:00, etc.), and the T arrive and T depart Compared with the cycle time boundary, if T arrive Falls in period C j If the time interval is within , the task is classified as C j After traversing all tasks, for each cycle C j The task time within the period is located at the extreme value, and the earliest arrival time T of all tasks within the period is extracted. earliest and the latest departure time T latest , generate replenishment time interval [T earliest ,T latest ], for example, a transport cycle C 001 It contains 3 replenishment tasks, and the time data is as follows:

[0091] Table 4 Example of replenishment task time classification

[0092]

[0093] As shown in Table 4, cycle C 001 The time interval is 08:00-12:00, traverse the arrival and departure time of tasks T001-T003, and extract T earliest =08:15 and T latest =11:45, but needs to be adjusted according to the cycle boundary rules: If the cycle start time is 08:00, then T earliest Force alignment to 08:00, if T latest If the end time of the cycle exceeds 12:00, it is split into multiple cycles. For example, the departure time of task T003 is 11:45, which is not beyond the limit, so cycle C 001 The final time interval is [08:00, 11:45], and the time interval length is calculated as ΔT = 11:45 - 08:00 = 3.75 hours.

[0094] S202: Based on the acquired replenishment time interval, the node congestion score is called to determine the overlap of time points between each replenishment time interval and the score of the corresponding node, and replenishment tasks with overlapping time periods are filtered and aggregated to obtain conflicting time intervals.

[0095] Based on the obtained replenishment time interval, the node congestion score data table (including Table 1) is called to extract the start time and end time of the replenishment task item by item, including the time interval of replenishment task A from 09:00 to 11:00 on April 18, 2025, and task B from 14:30 to 16:00. The congestion score sequence of node N1 from 09:00 to 12:00 is [85, 90, 95], and the score sequence of N2 from 14:00 to 17:00 is [70, 75, 80]. The time interval of the replenishment task is aligned with the time point of the node score minute by minute, including the time interval of task A from 09:00 to 11:00 and the time interval of N1 from 09:00 to 12:00 at 09: There is overlap between 00:00 and 11:00. The points per minute are compared to see if they exceed the preset threshold of 70. If the points at a single time point are ≥70, it is marked as a conflict. For example, if all points of Task A between 09:00 and 11:00 are ≥85, it is judged as a full-time conflict. If the points of Task B increase from 75 to 80 between 14:30 and 16:00, it is also judged as a full-time conflict. The final conflict time period is [09:00-11:00, 14:30-16:00].

[0096] Table 5 Node congestion score table

[0097] node Time interval Congestion Points N1 09：00-12：00 85-95 N2 14：00-17：00 70-80

[0098] As shown in Table 5, the congestion score range of node N1 from 09:00 to 12:00 is 85 to 95, and the score range of N2 from 14:00 to 17:00 is 70 to 80. The threshold is set at 70. According to data analysis, when the score is ≥70, the node traffic efficiency drops by more than 30%. By comparing the replenishment task time interval with the score interval, when the task time is completely included in the period with a score ≥70, it is determined to be a conflict. For example, 09:00-11:00 of task A is completely within the high score period of N1, and 14:30-16:00 of task B is completely within the high score period of N2. The screening of conflicting time periods is completed by minute-by-minute comparison, and finally a list of conflicting time periods is output.

[0099] S203: Based on the conflicting time interval, the replenishment time interval is called, and time axis shift processing is performed on the boundaries of both ends of the conflicting time interval. This includes adjusting the starting boundary time and the ending boundary time of the conflicting time interval as a shift operation to generate multiple non-overlapping new time intervals. The windows that intersect with the start and end times of the replenishment task are eliminated to obtain candidate rhythm intervals.

[0100] Based on the conflicting time intervals [09:00-11:00, 14:30-16:00], the original time interval 09:00-11:00 of replenishment task A and 14:30-16:00 of task B are called, and the time axis shift step is set to 30 minutes (based on the median calculation of the historical data sample [20, 25, 30, 35, 40] minutes, and the third digit value 30 is taken after sorting). The starting and ending ends of the conflicting time intervals are moved forward and backward respectively. For example, the starting end 09:00 of the conflicting time interval 09:00-11:00 is moved forward 30 minutes to 08:30, and the ending end 11:00 is moved backward 30 minutes to 11:30, generating new time intervals 08:30-10:30 and 09:30-11:30. The new time intervals are judged to overlap with the original replenishment time 09:00-11:00. The specific judgment criteria are: if the starting time t 新起始 Satisfy t 新起始 <t 原起始 And the new time period ends at time t 新结束 Satisfy t 新结束 ≤t 原起始 , then it is determined that there is no intersection, otherwise if t 新起始 ≥t 原起始 And t 新结束 ≤t 原结束, then it is determined that there is an intersection (for example, the start time 540 minutes of the new time period 09:30-11:30 ≥ the original task start time 540 minutes, and the end time 690 minutes ≤ the original task end time 660 minutes does not hold, so there is a partial intersection). After removing the windows with intersection, the new time period 08:30-10:30 without intersection is retained. Similarly, for the conflicting time period 14:30-16:00, it is moved forward 30 minutes to generate 14:00-15:30 and moved back 30 minutes to generate 15:00-16:30. The starting time 840 minutes of 14:00-15:30 is less than the original task starting time 870 minutes (14:30) and the ending time 930 minutes is less than the original task ending time 960 minutes (16:00), which is not true. It is determined that there is no intersection and this interval is retained. However, the starting time 900 minutes ≥ 870 minutes and the ending time 990 minutes ≤ 960 minutes of 15:00-16:30 is not true. There is an intersection and it is eliminated. Finally, the candidate rhythm intervals [08:30-10:30, 14:00-15:30] are obtained.

[0101] Table 6 Time axis displacement parameters and results

[0102]

[0103]

[0104] As shown in Table 6, the displacement step of 30 minutes is calculated from historical data samples. The specific steps are as follows: collect the adjacent feasible time window differences in the replenishment task adjustment records in the past three months. The record values ​​are [20, 25, 30, 35, 40] minutes, sort them and take the median of 30 minutes. The time overlap is determined by the minute digitization method. For example, 09:00 is converted to 540 minutes (9×60+0), and 11:00 is converted to 660 minutes. The start time of the new time period 08:30-10:30 is 510 minutes and the end time is 630 minutes. , compared with the original task time 540-660 minutes, if 510<540 and 630≤540 does not hold, but 510<540 and 630>540, it is determined to be a partial overlap. However, according to the correction logic (only when the new time period is completely outside the original task time), since the end time of the new time period 630 minutes (10:30) is earlier than the original task start time 540 minutes (09:00), it does not hold, so it is corrected to allow partial non-inclusive overlap, and only completely inclusive or overlapping windows are eliminated. Finally, through the interval inclusion relationship operation of the minute value (t 新起始 ≥t 原结束 or t 新结束 ≤t 原起始) Filter out non-intersecting candidate intervals. For example, the interval 840-930 minutes from 14:00 to 15:30 satisfies the original task time window of 870-960 minutes, but the interval 840<870 and 930≤870 does not hold. However, 840<870 and 930>870 are still considered non-intersecting according to the modified logic because it does not completely cover the original task time window. Finally, a list of rhythm candidate intervals is output.

[0105] S204: Based on the obtained rhythm candidate intervals, all candidate intervals are sorted in ascending order according to the start time, and time window combinations that do not overlap within a single cycle are selected to generate a replenishment rhythm window;

[0106] Based on the candidate rhythm intervals [08:30-10:30, 14:00-15:30], call the start time parameters of all candidate intervals (for example, 08:30 is converted to the minute value 510, 14:00 is converted to 840), and sort them in ascending order according to the numerical value of the start time to generate a sorted list [510-630 (08:30-10:30), 840-930 (14:00-15:30)], traverse each time period in the list, starting from the first time period, and compare the end time t of the current time period. 当前结束 The start time of the next time period is t 下一起始 , if t 当前结束 +Δt 阈值 ≤t 下一起始 (where Δt 阈值 For example, if the preset time interval threshold is 15 minutes and the shortest passage interval of the warehouse forklift is used as the basis, it is determined to be non-overlapping and the next time period is retained. Otherwise, it is eliminated. For example, comparing 510-630 and 840-930, calculating 630+15=645≤840, the condition is met and both are retained. If there is a third time period 960-1020 (16:00-17:00) in the list, then comparing 930+15=945≤960, the condition is met and retained. If there is an overlapping time period such as 840-930 and 900-1000, the calculation of 930+15=945≤900 does not hold, then 900-1000 is eliminated, and finally a non-overlapping time window combination [08:30-10:30, 14:00-15:30] is generated.

[0107] Table 7 Candidate interval ranking and screening results

[0108] Candidate time period Start time (minutes) End time (minutes) Whether to retain 08:30-10:30 510 630 yes 14:00-15:30 840 930 yes 15:45-17:00 945 1020 yes

[0109] As shown in Table 7, the time interval threshold Δt 阈值It is set to 15 minutes. It is obtained by counting the minimum buffer time required for the warehouse forklift to leave the node after loading and unloading and to start the next task. The specific operation is to collect 50 forklift task interval data in the past week, record the sample [10, 12, 15, 18, 20] minutes, and take the 80% quantile 15 minutes. When screening, starting from the first time period after sorting, check whether the interval between adjacent time periods is ≥15 minutes. For example, the interval between the first time period 08:30-10:30 (510-630) and the second time period 14:00-15:30 (840-930) is 840- 630=210 minutes ≥ 15 minutes, retained. If there is a third time period 15:45-17:00 (945-1020), check that the interval between it and the second time period is 945-930=15 minutes, which meets the threshold and is retained. If the interval between a time period and the previous time period is <15 minutes, for example, 14:00-15:30 (840-930) and 15:00-16:00 (900-960), then calculate 930+15=945≤900, which does not hold, and is determined to be crossed. 15:00-16:00 is eliminated, and a list of non-crossing time window combinations is finally output.

[0110] See also Figure 1 The replenishment estimated arrival start time and the replenishment estimated arrival end time in the replenishment rhythm window are matched with the warehouse-side shipping schedule. The warehouse-side picking and delivery sequence, the transportation channel usage sequence, and the shipping schedule data are serialized and sorted. The specific steps for generating the path segment sequence code include:

[0111] S301: Obtain the estimated arrival start time and the estimated arrival end time of replenishment in the replenishment rhythm window, match the start and end times of the replenishment tasks in the replenishment rhythm window with the scheduled shipping time at the warehouse end, call the warehouse task node information to perform logical judgment on the matching group, group the task combinations with overlapping time into the same rhythm group, and generate the time matching results between the replenishment tasks and the warehouse nodes.

[0112] Based on the replenishment rhythm windows 08:30-10:30, 14:00-15:30, extract the estimated arrival start time t of replenishment task A 补货起始 =08:30 (converted to minute value t 分钟 =8×60+30=510) and the end time t 补货结束 = 10:30 (630 minutes), t for Task B 补货起始 =14:00(840 minutes) and t 补货结束 = 15:30 (930 minutes), call the shipping schedule window t of storage node N1 仓储起始 =09:00 (540 minutes) and t 仓储结束= 12:00 (720 minutes), and t 仓储起始 =13:00 (780 minutes) and t 仓储结束 = 16:00 (960 minutes). After quantifying the replenishment task time and the storage node time in minutes, perform the following logical judgment item by item.

[0113] 1. Time overlap determination: For each pair of replenishment tasks and storage nodes, check whether (t 补货起始 <t 仓储结束 )∧(t 补货结束 >t 仓储起始 ), for example, if Task A (510-630 minutes) and N1 (540-720 minutes) are combined, the calculation is 510 < 720 and 630 > 540. If the conditions are met, they are grouped together in the same rhythm group. If Task B (840-930 minutes) and N2 (780-960 minutes) are combined, the calculation is 840 < 960 and 930 > 780. If the conditions are met, they are grouped together in the same rhythm group. 2. Non-overlapping task processing: If the time window of replenishment task C is 08:00-08:30 (480-510 minutes) and N1 (540-720 minutes), the calculation is 510 ≤ 540. It is determined to be non-overlapping and grouped separately. 3. Multi-node priority allocation: If the replenishment task overlaps with multiple nodes at the same time (for example, Task A overlaps with both N1 and N3), the node load factor is calculated. For example, L of N1 节点 =3 / 5=0.6, L of N3 节点 =4 / 5=0.8, select L 节点 The smaller N1 is aggregated.

[0114] Table 8 Time matching judgment results

[0115] Restocking tasks Replenishment time (minutes) Storage Node Storage time (minutes) Overlap A 510-630 N1 540-720 yes B 840-930 N2 780-960 yes C 480-510 N1 540-720 no

[0116] As shown in Table 8, the time overlap determination formula is:

[0117] Overlap condition = (t 补货起始 <t 仓储结束 )∧(t 补货结束 >t 仓储起始 ),

[0118] where t 补货起始 , t 补货结束 , t 仓储起始 , t 仓储结束 All through t 分钟 = hours × 60 + minutes conversion, for example, 09:00 of N1 is converted to 540 minutes, and the node load factor L 节点By dividing the number of real-time tasks by the preset maximum load (e.g. N1 can carry up to 5 tasks), the rhythm group list is finally generated: Group 1: A-N1, Group 2: B-N2, Group 3: C-Independent.

[0119] S302: Based on the obtained time matching results between the replenishment tasks and the storage nodes, the warehouse picking plan number, transportation channel code, and shipping appointment time corresponding to the tasks in the rhythm group are called, and the three types of data are bound to perform task association sequence. The parameters in the task group are jointly sorted according to the time sequence relationship to obtain the operation rhythm sorting sequence of the associated nodes after sorting;

[0120] According to the time matching result between the replenishment task and the storage node (for example, the arrival time of the replenishment task is 2025-04-18 10:00:00, the warehousing cut-off time of storage node A is 10:30:00, and that of node B is 11:00:00), the picking plan number corresponding to the task in the rhythm group is called (for example, task group T001 contains picking plans P001 and P002), and the transportation channel code associated with the picking plan is retrieved from the warehouse management system (for example, channel C01 corresponds to P001, and C02 corresponds to P002). The picking plan number, transportation channel code and shipping appointment time (for example, the shipping time of P001 is 10:15) are matched. 00, P002 is 10:45:00) are arranged in chronological order. The specific operation is as follows: extract the hour and minute values ​​of the shipping appointment time (10:15 is converted to the value 1015, and 10:45 is converted to 1045), and arrange the picking plan number and channel code in ascending order according to the value (P001-C01 is arranged before P002-C02). If the time is the same, arrange them in ascending order according to the last digit of the channel code (for example, C01 and C03, C01 takes priority), and finally generate the sorting sequence (T001-P001-C01-10:15:00, T001-P002-C02-10:45:00).

[0121] Table 9 sorting sequence:

[0122] Storage node code Replenishment task arrival time Warehousing deadline Number of available channels WH-A 2025-04-18 10:00:00 2025-04-18 10:30:00 2 WH-B 2025-04-18 10:00:00 2025-04-18 11:00:00 3

[0123] As shown in Table 9, the difference between the arrival time of the replenishment task and the storage node's inbound deadline is used as the judgment basis (the time difference is 30 minutes for WH-A and 60 minutes for WH-B). If the time difference is ≤15 minutes (the threshold is set according to the "Emergency Task Processing Time Limit" clause in the storage node operation manual), the node with ≥2 available channels is prioritized (WH-A meets the condition). Otherwise, the node with a larger time difference is selected (if the time difference between WH-B exceeds the threshold, manual review is required). In this example, the task arrival time is 10:00:00, and the inbound deadline of WH-A is 10:30:00. The calculated time difference is 30 minutes (exceeding the threshold of 15 minutes), so channel availability is checked: WH-A currently has 2 available channels (meeting the ≥2 condition), so it is assigned to WH-A. Otherwise, a backup node is called.

[0124] The ordering of shipping appointment times is as follows: convert the times to 4-digit numbers (e.g., 10:15 → 10:15, 10:30 → 10:30), sort them in ascending order, and if the numbers are the same, sort them in ascending order by the last digit of the channel code (e.g., C01 and C02, C01 comes first). Task group T001 contains two shipping times (10:15 and 10:45), which after conversion are 10:15 and 10:45, respectively. They are sorted in ascending order from 10:15 to 10:45, with the corresponding associated parameters being P001-C01 and P002-C02. If there are tasks with the same time (e.g., both with 10:15), they are further sorted by the last digit of the channel code (e.g., C01 and C03, C01 comes first).

[0125] The resulting job rhythm sorting sequence is: T001-P001-C01-10:15:00, T001-P002-C02-10:45:00. The logic is that time difference takes precedence over channel availability, and the last digit of the channel code is the secondary sorting basis.

[0126] S303: Based on the generated operation rhythm sorting sequence, the warehouse number, channel number, and shipping number in the sorting sequence are called and sequentially combined to generate path segment structure information. The path fields within the same rhythm group are marked with unique picking sorting codes, and sequence segment numbers are established and the corresponding task-path mapping relationship is recorded to generate a path segment sequence code.

[0127] According to the operation rhythm sorting sequence (for example, the sequence includes warehouse number WH-A, channel number C01, and shipment number S001, which are arranged in chronological order as WH-A→C01→S001), the warehouse number (WH-A), channel number (C01), and shipment number (S001) are extracted, and the three are spliced ​​into path segment structure information (WH-A-C01-S001) in the format of "warehouse-channel-shipment". The path fields in the same rhythm group (for example, rhythm group T001 includes WH-A-C01-S001 and WH-A-C02-S002) are sorted. Row uniqueness check: Extract the last two characters of the warehouse number (A), the last two digits of the channel number (01), and the last three digits of the shipment number (001), and generate a unique code (A-01-001) according to the "last digit of the warehouse + last digit of the channel + last digit of the shipment" rule. If the code is repeated (such as A-01-001 already exists), append a sequential identifier (A-01-001-1) to the end, establish a serial segment number (starting from 001 and increasing in the order of generation), bind the serial segment number to the path segment code (001 corresponds to A-01-001), and store it in the task path mapping relationship table.

[0128] Table 10 Task path mapping relationship table:

[0129]

[0130] As shown in Table 10, the path segment code generation rule is as follows: truncate the last letter of the warehouse number (e.g., A for WH-A), truncate the last two digits of the channel number (01 for C01), and truncate the last three digits of the shipment number (001 for S001), concatenating them to form A-01-001. If the same code exists within the same rhythm group (e.g., two A-01-001s), append a sequence identifier (A-01-001-1). In this example, rhythm group T001 contains channels C01 and C02 of warehouse WH-A, with shipment numbers S001 and S002, respectively. Path segment codes A-01-001 and A-02-002 are generated, and the sequence segment numbers are assigned to 001 and 002 in the order they are generated.

[0131] The specific operation for uniqueness verification is: call the path segment encoding library, compare the newly generated code (A-01-001) with the records in the library one by one, and if there is an exact match, append "-" and the current time minute value at the end (for example, if the current time is 10:15, append -15) to generate A-01-001-15. If there is still a duplicate (for example, A-01-001-15 already exists in the library), continue to append the seconds (append -30 to 10:15:30) until it is unique. Assuming A-01-001 already exists in the library, the new code is appended with -15, generating A-01-001-15. If another task simultaneously generates the same code, the number of seconds is further appended (e.g., A-01-001-15-30). The rule for assigning sequential segment numbers is as follows: starting with 001, the segment number increases by 1 for each new path segment code. If the same rhythm group contains multiple path segments (e.g., T001 contains A-01-001 and A-02-002), consecutive segment numbers (001 and 002) are assigned based on the order in which the path segment codes were generated. In this example, channel C03 of warehouse WH-B and shipment S003 generate path segment code B-03-003, with segment number 003.

[0132] The task path mapping relationship is recorded by storing the sequence segment number (001), path segment code (A-01-001), warehouse number (WH-A), channel number (C01), and shipping number (S001) in a database table, with foreign keys linking the fields. If a path segment code is modified due to a conflict (e.g., A-01-001-15), the code field in the mapping table is updated synchronously, and the original code is retained as a record. The resulting path segment sequence code is: 001-A-01-001, 002-A-02-002, 003-B-03-003. The logic is to bind the unique code and sequence segment number in chronological order, ensuring the traceability of the task path.

[0133] See also Figure 1 ,The specific steps of calling the path segment sequence code to calculate the similarity of the vehicle path sequence records collected by the vehicle GPS, screening the vehicle resources with similar sequences and binding the corresponding vehicle to the path label include:

[0134] S401: Calling the path segment sequence code, obtaining the vehicle path sequence record collected by the vehicle GPS, comparing the matching degree of the three fields of the path segment, namely the numbered node, time sequence and channel number, calculating the structural coincidence rate between the current path segment and the preset reference value of the vehicle path segment, and generating the path fit;

[0135] Call the vehicle path sequence record R collected by the vehicle GPS vehicle ={(n1,t1,c1),(n2,t2,c2),…,(n p ,t p ,cp )}, where n i is the node number of the path segment, t i is the timestamp, c i is the channel number, extract the current path segment R current ={(n′1,t′1,c′1),(n′2,t′2,c′2),…,(n′ q ,t′ q ,c′ q )}, traverse R vehicle For each vehicle's path segment, the node number, time sequence, and channel number fields are matched item by item. For example, the path segment of vehicle A is {(101, 08:00, 5), (102, 08:05, 5), (103, 08:10, 5)}, and the current path segment is {(101, 08:02, 5), (102, 08:07, 5)}. The number of node number matches is 2 (101, 102). The time sequence difference is calculated as |08:02-08:00| = 2 minutes and |08:07-08:05| = 2 minutes. The average minutes, the number of channel number matches is 2 (all 5), and the path fitting degree S fit =0.94×100=94, set the path matching threshold to 85%, and filter S fit For vehicles with a speed of ≥85, vehicle B's path segment is {(101, 08:15, 6), (104, 08:20, 6)}, and the number of node matches with the current path segment is 1 (only 101), the time sequence difference |08:02-08:15| = 13 minutes (exceeding the tolerance of 10 minutes), and the number of channel number matches is 0. S fit =30, is excluded. Table 11 lists the vehicle path matching data.

[0136] Table 11 Vehicle path matching data table

[0137]

[0138] As shown in Table 11, the S fit All of them are higher than 85% and are added to the candidate set. Data analysis shows that for every 10% reduction in node matching error, task efficiency increases by 15%, while a 10% reduction in time sequence error only increases efficiency by 3%, and a 10% reduction in channel number error increases efficiency by 1%. By allocating weights through linear regression, the tolerance threshold for time sequence difference is set to 10 minutes. The reason is that a time deviation of more than 10 minutes causes a 25% increase in the task scheduling conflict rate. For example, the time difference of vehicle D is 12 minutes, and the calculation result of item w2 is (0 is assigned directly due to over-limit). In the channel number matching judgment, if the vehicle path is inconsistent with the current path channel (such as vehicle B's channel 6 vs the current channel 5), it is directly counted as 0 matching. The threshold of 85% is set based on the task success rate statistics: S fit ≥85 mission success rate is 92%, S fit When the success rate is less than 85%, it drops to 65%. By adjusting the threshold to 85%, vehicle A (S fit =94) and C(S fit =89), the three types of matching proportions were directly added together and the average was taken.

[0139]

[0140] Calculate the structural overlap rate S overlap ;

[0141] Among them, N match It is the node number set corresponding to the real-time data of the vehicle's current path status and the preset benchmark of the vehicle path segment. The number of node number matches formed by counting the number of identical numbers is T. max is the time difference tolerance threshold, which is obtained by statistical quantiles, N total Indicates the total number of nodes in the current path segment. represents the mean of the time sequence difference, C match Indicates the number of channel number matches, C total Indicates the total number of channels in the current path segment.

[0142] S402: Based on the acquired path fit, the idle time and available loading status parameters of each vehicle in the vehicle status list in the current cycle are retrieved, a path matching threshold is determined based on the path overlap rate, and vehicles with path fits higher than the path matching threshold and meeting the status conditions are selected as candidate sets to establish vehicle matching resources.

[0143] According to the current path coordinate point sequence P current ={(x1,y1),(x2,y2),…,(x n ,y n )} and the vehicle path coordinate point sequence P vehicle ={(x′1,y′1),(x′2,y′2),…,(x′ m ,y′ m )}, calculate the path fitting degree S fit The specific process is: extract P current With P vehicle For the first k points (k = min (n, m)), calculate the lateral distance deviation Δd for each point i =|x i -x′ i |+|y i-y′ i |, sum and take the mean Will Mapped to (like Then S fit =0), for example, the Δd between a vehicle path and the first five points of the current path i They are 0.8, 1.2, 0.5, 1.0 and 0.7 meters respectively. m, then S fit =100-8.4=91.6, the path matching threshold is set to 85, and by comparing the idle time T in the vehicle status list idle (Unit: minute) and available load capacity L load (Unit: Percentage), Screening T idle ≥30 and L load ≥80% of vehicles, in this example, the T of vehicle A idle =45,L load =90%, S fit =91.6, meeting the conditions and being added to the candidate set. Vehicle B’s T idle =25,L load =85%, S fit =88, because T idle <30 are excluded, and Table 12 lists the candidate vehicle state parameters.

[0144] Table 12 Candidate vehicle status parameters

[0145] Vehicle number Job idle time (minutes) Available loading capacity (%) Path Fit A 45 90 91.6 C 35 95 86.2 D 40 85 87.5

[0146] As shown in Table 12, the S fit All above 85, and T idle ≥30, L load ≥80%, forming vehicle matching resources.

[0147] Path matching threshold S threshold =85 is set based on the following: the average delay rate of vehicles with a path fit of ≥85 in the statistical task is 5%, while S fit The average delay rate of vehicles <85 is 22%. threshold To control the delay rate within 10%, for example, when S threshold =80, the delay rate is 15%, S threshold = 85, the delay rate is reduced to 8%, and 85 is finally selected as the threshold. The job idle time threshold T threshold =30 minutes is based on the average loading and unloading time of 25 minutes, with a 5-minute tolerance. For example, if a task is expected to take 28 minutes, and vehicle T idle= 30, then the remaining time is 30-28 = 2 minutes, which needs to be excluded. However, in practice, due to T threshold If the vehicle is still selected, the system will automatically assign a vehicle with a higher priority. The available loading capacity threshold L threshold =80% is set based on: when L load When the L of vehicle E is less than 80%, the vehicle needs to be dispatched to replenish the goods, which increases the task time by 18%. load =75%, path fit 89, due to L load <80% were excluded.

[0148] S403: Based on the acquired vehicle matching resources, the path matching threshold is determined by calculating the distribution characteristics using the time difference data. The unique identification code of each candidate vehicle and the corresponding path segment sequence code are used to generate a label combination. The combined labels are grouped by vehicle and written into the vehicle task assignment field to establish a vehicle deployment path label.

[0149] Extract the unique identification code V of the candidate vehicle from the vehicle matching resource id (For example, V id =[VH001, VH002, VH003]), call the path segment sequence code P corresponding to each vehicle code (For example, P code =[PS001-202404,PS002-202404,PS003-202404]), V id With P code Press "V id -P code " format to generate a combined label. In this example, the path segment code of vehicle VH001 is PS001-202404, and the label "VH001-PS001-202404" is generated. All candidate vehicles are traversed. If a vehicle has multiple path segments (such as vehicle VH003 associated with PS003-202404 and PS004-202404), multiple labels are generated and merged into a comma-separated string "VH003-PS003-202404, VH003-PS004-202404". When grouping, press V id Index grouping, for example, the label of vehicle VH002 is "VH002-PS002-202404", which is written into the database task assignment field F task Table 13 shows the label collection results of the three candidate vehicles.

[0150] Table 13 Vehicle task assignment fields and path label table

[0151]

[0152] As shown in Table 13, vehicle VH003 contains two path segment codes, so two labels are generated and separated by commas. In the label generation rule, the unique identification code V id The format is "VH" + 4 digits (such as VH001 to VH999), the path segment code P code The format is "PS" + 3 digits + "-" + 6 digits of year and month (such as PS001-202404 represents the first path segment of April 2024). The original encoding characters are strictly retained when combining, and truncation or escape is prohibited. For example, in the path segment PS002-202404, "202404" is the year and month identifier and is directly spliced ​​into part of the label. During the collection process, the system will use V id Hash value is used to establish an index and locate the vehicle task assignment field with O(1) time complexity. For example, the hash value of vehicle VH001 is obtained by the hash function H(V id )=H("VH001")=0x3A7D is mapped to memory address 0x3A7D and the label data is written directly. For vehicles with multiple path segments, the labels are arranged in ascending order by timestamp. For example, the timestamp of the path segment PS003-202404 of vehicle VH003 is 2024-04-2008:00:00, and the timestamp of PS004-202404 is 2024-04-2008:30:00. The labels are arranged in the order of "PS003-202404, PS004-202404". The path label is written to field F task When checking whether the field length exceeds 256 characters, if it exceeds the limit, the compression algorithm is triggered. For example, vehicle VH004 contains 10 path segments, the original label length is 215 characters, which is within the limit and is written directly. Vehicle VH005 contains 15 path segments, the label length is 310 characters, and the LZW compression algorithm is enabled. The compressed length is 172 characters. After writing is completed and the vehicle delivery path label is established, the system will F task The fields are synchronized to the dispatch center. For example, the label "VH001-PS001-202404" of vehicle VH001 is parsed as a delivery instruction: assign path segment PS001-202404 to VH001.

[0153] See also Figure 1 Based on the replenishment rhythm window matching judgment, the delivery instructions are sent, the consistency between the delivery path and the vehicle placement path label is monitored in real time, and the specific steps of generating the transportation scheduling adjustment result according to the monitoring results include:

[0154] S501: Obtain the arrival start and end times of the replenishment task in the replenishment rhythm window, call the vehicle delivery path tag, match the start and end times of the replenishment task with the path segment times recorded in the corresponding vehicle path tag, and measure the time difference to generate a rhythm matching offset result for the current path task;

[0155] Extract the arrival time range T of the replenishment task from the replenishment rhythm window task =[t start ,t end ](e.g. t start =08:00:00, t end = 09:30:00), parsing the path segment time T in the vehicle placement path label path =[t p1_start ,t p1_end ,t p2_start ,t p2_end ,…] (e.g. the path label “VH001-PS001-202404” of vehicle VH001 corresponds to the time of path segment PS001 as [08:05:00,08:55:00]), and the time difference Δt is calculated for each path segment. start =|t p_start -t start | and Δt end =|t p_end -t end |, if the path segment is multiple (such as PS003 and PS004 of vehicle VH003), then match each segment and take the average value. For example, the PS003 time of VH003 is [08:10:00, 08:50:00], and the PS004 time is [09:00:00, 09:40:00], Δt start1 =10 minutes, Δt end1 =40 minutes, Δt start2 =60 minutes, Δt end2 =10 minutes, average Total offset O total =0.6×35+0.4×25=31 minutes (time difference weight w start =0.6, w end =0.4), set the offset threshold O threshold = 30 minutes, if O total ≤O threshold The match is considered successful. In the example, the O of VH003 total =31 exceeds the threshold and is excluded. The path segment time of vehicle VH002 is [08:15:00, 09:20:00], Δt start =15,Δt end =10,O total =0.6×15+0.4×10=13, which is lower than the threshold and the matching is successful. Table 14 shows the offset calculation results of the three vehicles.

[0156] Table 14 Replenishment task time and path segment time offset table

[0157]

[0158] As shown in Table 14, the O total All are below the threshold of 30 and are marked as matching success, with a weight of w start =0.6 and w end =0.4 is set based on the following reason: the influence of the starting time deviation of the replenishment task on the warehouse stocking preparation is more important. Data analysis shows that Δt start For every 10 minutes increase in delay, the replenishment delay rate increases by 12%, while Δt end An increase of 10 minutes only results in a 5% increase in delay rate. By assigning weights through linear regression, the offset threshold O threshold =30 minutes is based on actual operational tolerance: a time deviation exceeding 30 minutes will increase the subsequent vehicle scheduling conflict rate by 40%, for example, the O total =32, the conflict rate statistics are 42%, so the threshold is set to 30. In the time difference calculation, if the route segment spans multiple replenishment windows (for example, the route segment of vehicle VH005 is [07:50:00, 09:00:00]), only the time period overlapping with the replenishment window is matched. For example, if the overlapping time period is [08:00:00, 09:00:00], calculate Δt start =0,Δt end =30,O total =0.6×0+0.4×30=12, the match is successful. For non-integer time data (such as t start = 08:12:37), convert it to decimal format of minutes (08:12:37 = 8 × 60 + 12 + 37 / 60 = 492.62 minutes), and then use it in the calculation. For example, the path segment start time 08:10:00 is converted to 490 minutes, Δt start =|492.62-490|=2.62 minutes, which is 2.62 after rounding to two decimal places.

[0159] S502: Based on the obtained rhythm matching offset result, the route task numbers with time offsets less than the offset threshold are screened, the corresponding vehicle numbers and replenishment task numbers are called to execute the combinations, a delivery instruction is generated for each combination, the route number and issuance time are recorded, and a delivery task is established;

[0160] Screening O total Path tasks with a value of ≤30 (such as PT001 and PT002) will correspond to V id (VH001, VH002) and R task (RT001, RT002) spliced ​​into C id (VH001-RT001, VH002-RT002), generate delivery instruction D cmd , including Pcode (PS001-202404), G type (Electronic products), L load (90%), record T issue (2024-04-2009:00:00), Table 15 shows the three sets of task data.

[0161] Table 15 Delivery task instruction table

[0162] Vehicle number Replenishment task number Path number Issue time VH001 RT001 PS001-202404 2024-04-20 09:00:00 VH002 RT002 PS002-202404 2024-04-20 09:05:00 VH004 RT004 PS005-202404 2024-04-20 09:10:00

[0163] As shown in Table 15, C id By V id With R task Generate by connecting with a short dash, P code Strictly consistent with the path label, T issue Taken from the system clock, if the path number is repeated (such as PS005-202404 already exists), add the "-n" suffix (such as PS005-202404-1) and write F deliver Fields are checked for uniqueness. For example, if the path PS001-202404 of vehicle VH007 is repeated, PS001-202404-2 is generated. The formula P code_new =P code +, -, +(n+1), n ​​is the existing number of repetitions, n=1, and P is obtained. code_new =PS001-202404-2.

[0164] S503: Call the obtained delivery task, collect the current positioning trajectory of the delivery vehicle and compare it with the preset path segment sequence number, extract the current vehicle operation number of each path segment and the delivery tag path segment number to determine the node consistency, obtain the ratio of the number of consistent nodes in each path segment to the total number, and generate the path consistency ratio;

[0165] Call delivery task field F deliver Path number P in code (such as PS001-202404), collect the real-time positioning trajectory of the vehicle L real ={n1,n2,…,n m} (e.g. n1=101, n2=102, n3=104), extract the preset path segment sequence P preset ={n′1,n′2,n′3}(e.g. PS001-202404 corresponds to n′1=101,n′2=102,n′3=103), traverse each path segment, and compare L node by node real With P presetFor example, in path segment 1, n1=101 matches n′1=101, n2=102 matches n′2=102, and n3=104 does not match n′3=103. The matching number N match =2, total number of nodes N total =3, node consistency ratio R node =2 / 3≈66.7%, synchronous calculation timestamp deviation Δt i =|t real_i -t preset_i |(such as t preset_1 =08:05:00, t real_1 = 08:06:30, Δt1 = 90 seconds), if Δt i ≤T threshold = 300 seconds (5 minutes), it is counted as time matching, and the time matching number T match =2 (assuming that the time deviation of the two nodes is qualified), the total ratio R time =2 / 3≈66.7%, comprehensive consistency ratio R consist =0.7×R node +0.3×R time =0.7×0.667+0.3×0.667=0.667(weight w node =0.7, w time =0.3), set the consistency threshold R threshold =0.65, determine R consist =0.667≥0.65 is qualified. Table 16 shows the matching data of the three vehicles.

[0166] Table 16 Path node and time consistency matching table

[0167]

[0168]

[0169] As shown in Table 16, the R consist Above the threshold 0.65, the node matching weight w node =0.7 is set based on the following: Historical data statistics show that when the node matching error is reduced by 10%, the delivery accuracy rate increases by 18%, while a 10% reduction in the time matching error only increases the accuracy rate by 5%. By allocating weights through linear regression, the time deviation threshold T threshold =300 seconds is set based on the actual road condition tolerance: a time deviation of more than 5 minutes will increase the conflict rate of subsequent tasks by 35%. For example, the vehicle VH005 has Δt = 310 seconds, and the conflict rate statistics are 38%. In the path segment node matching judgment, if the actual trajectory node of the vehicle is completely consistent with the preset node number (such as 3 / 3 matching of VH002), then Rnode =1.0. If there is a missing or offset (such as 1 / 2 match of VH004), the actual ratio is calculated. R for non-integer values consist Keep three decimal places (such as 0.667). For tasks that span multiple path segments (such as the path of vehicle VH006 includes PS003 and PS004), calculate R for each segment separately. consist Then take the mean, for example, R of PS003 consist =0.7, R of PS004 consist =0.8, final R consist =(0.7+0.8) / 2=0.75.

[0170] Formula Description:

[0171] The comprehensive consistency ratio formula is:

[0172]

[0173] N match : The number of matching between actual trajectory nodes and preset node numbers, by traversing L real With P preset A set of nodes, counting the number of nodes with the same number, such as N for VH001 match =2, N total : The total number of nodes in the preset path segment, for example N total =3, T match : Time deviation Δt i The number of nodes with a timeout of ≤300 seconds, such as T for VH001 match =2, T total =N total :The total number of nodes is consistent with the total number of time verification, w node =0.7, w time =0.3.

[0174] The path segment PS003-202404 of vehicle VH003 has the preset nodes {201, 202, 203}, the actual trajectory nodes are {201, 202, 204}, and the time deviation is {60, 120, 400} seconds. The calculation process is as follows:

[0175] 1.N match =2 (matching 201 and 202), R node =2 / 3≈0.667,2.T match =2(Δt1=60≤300, Δt2=120≤300, Δt3=400>300), R time =2 / 3≈0.667,3.R consist =0.7×0.667+0.3×0.667=0.667, 4. Comparison threshold R threshold=0.65, since 0.667≥0.65, it is judged to be qualified.

[0176] The results show that the path segment PS003-202404 meets the consistency requirements. The formula uses weighted comprehensive node and time matching to avoid bias in the results caused by a single factor.

[0177] S504: Based on the acquired path consistency ratio, the path segment sequence code is called, and the path segment number is replaced based on the real-time scheduling deviation of the path deviation benchmark ratio. The replacement path segment deviation result is recorded and appended to the control scheduling log to generate the transportation scheduling adjustment result.

[0178] Extract path consistency ratio R consist (including R consist =[0.667,0.500,0.750]) and the path deviation reference ratio R base =0.65, screening R consist <R base Vehicle number V id (Including VH004's R consist =0.500), call its original path segment number P code (including PS005-202404), from the path pool P pool Select the new path segment P new (including PS006-202404), the selection rule is: prioritize the matching node number difference ΔN=|N new -N old |≤2(including N old =3,N new =4, ΔN=1), and the time coverage range T new Overlap rate with replenishment window η overlap ≥80%, including the original path PS005-202404 time range of [09:], the new path PS006-202404 is [09:], the overlapping time is 45 minutes, the total time is 60 minutes, η overlap =45 / 60=75% (does not meet the conditions), continue to screen PS007-202404 ([09:], overlap 55 minutes, η overlap =55 / 60≈91.7%), replace after meeting the conditions, and record the replacement operation to the scheduling log L log , the field includes the original path P code 、New path P new , replacement time T replace (including 2024-04-2010:00:00), Table 17 shows the replacement results of the three vehicles.

[0179] Table 17 Path segment replacement record table

[0180] Vehicle number Original path segment number New path segment number Replacement time VH004 PS005-202404 PS007-202404 2024-04-20 10：00：00 VH006 PS008-202404 PS009-202404 2024-04-20 10：05：00 VH007 PS010-202404 PS011-202404 2024-04-20 10：10：00

[0181] As shown in Table 17, the original path PS005-202404 of vehicle VH004 is due to R consist =0.500<0.65 is replaced by PS007-202404, the base ratio R base =0.65 is set based on historical data analysis: when R consist <0.65, the delivery error rate rises to 25%, while R consist When the error rate is less than 10% when ≥0.65, the path pool P pool The node number difference limit ΔN≤2 is calculated by the actual path length, including the original path length L old = 15 km, the new path length is L new =17 km, ΔL = 2 km corresponds to ΔN = 1 (about 1 node every 2 km), the overlap rate threshold η overlap ≥80% is set to avoid the replenishment time window being cut, including the original window is 09:00-10:00. If the new route time is 09:30-10:30, the overlap rate is only 50%, which does not meet the conditions. When the route is replaced, the system checks P new The node sequence and the vehicle's current position L current Connectivity, including vehicle VH004 at node 105, the new path PS007-202404 must include 105 or adjacent node 106. If it is not connected (including PS007-202404 starting node 107), then trigger the secondary screening and replace the record written to L log The JSON format is:

[0182] {V id : "VH004", P code : "PS005-202404", P new : "PS007-202404", T replace : "2024-04-2010:00:00"};

[0183] Press V when the log is saved id Hash shards, including the hash value H(V id )=0x5A3E corresponds to shard 3, ensuring query efficiency.

[0184] Formula Description:

[0185] The overlap rate calculation formula is:

[0186]

[0187] -t start , tend : Replenishment task time window, -T new_start 、T new_end : Time range of the new route segment.

[0188] Replenishment Window start =09:00:00 (540 minutes), t end =10:00:00 (600 minutes), new path time T new_start =09:15:00 (555 minutes), T new_end =10:05:00 (605 minutes), calculate:

[0189] 1. Overlap start time max(540,555)=555, 2. Overlap end time min(600,605)=600, 3. Overlap duration 600-555=45 minutes, 4. Total duration of the new path 605-555=50 minutes, 5. η overlap =45 / 50×100%=90%.

[0190] This result shows that the new path meets the overlap rate threshold. The formula ensures the timeliness of tasks after path replacement by accurately calculating the time intersection ratio.

[0191] See also Figure 2 A logistics and transportation management system based on multi-source data fusion is provided. The logistics and transportation management system based on multi-source data fusion is used to implement the above-mentioned logistics and transportation management method based on multi-source data fusion. The system includes:

[0192] The node load integration module monitors the transport time of logistics vehicles in real time, inputs the vehicle transport time into the principal component analysis model, calculates the congestion score of the transport route, supplements the node load integration data, and transmits the generated congestion score to the replenishment rhythm window module;

[0193] The replenishment rhythm window module obtains the estimated replenishment arrival start and end times of the replenishment task in the next transportation cycle, determines the overlapping replenishment time periods based on the congestion score of the transportation route, performs shifting based on the start and end times to generate multiple rhythm window candidate time periods, and passes them to the path segment sequence generation module;

[0194] The path segment sequence generation module matches the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse-side shipping scheduled time, and serializes and sorts the warehouse-side picking and outbound order, the transportation channel usage order, and the shipping scheduled time data to generate a path segment sequence code and pass it to the vehicle label binding module;

[0195] The vehicle label binding module calls the current path segment sequence in the path segment sequence code, compares the fit with the path segment sequence of the previous period, selects vehicle resources with similar sequences, binds the corresponding vehicle delivery path labels, and transmits them to the collaborative scheduling control module;

[0196] The collaborative scheduling control module sends delivery instructions based on the replenishment rhythm window matching judgment, monitors the consistency of the delivery path and the vehicle delivery path label in real time, and generates transportation scheduling adjustment results based on the monitoring results.

[0197] It should be understood that the term "and / or" as used herein simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the related objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0198] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0199] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0200] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein 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. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0201] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0202] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0203] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0204] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0205] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0206] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A logistics and transportation management method based on multi-source data fusion, characterized in that: The following steps are involved: S1: Real-time monitoring of the transport time of logistics vehicles, inputting the transport time of the vehicles into the principal component analysis model, and calculating and recording the congestion score of the transport route; S2: Obtain the estimated arrival start and end times of the replenishment tasks within the next transportation cycle, determine the overlapping replenishment time periods based on the congestion score of the transportation route, perform shifting based on the start and end times to generate multiple rhythm window candidate time periods, and determine the replenishment rhythm window; S3: Matching the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse-side scheduled shipping time, sorting the warehouse-side picking and delivery sequence, transportation channel usage sequence, and scheduled shipping time data using a dynamic priority scheduling algorithm to generate a path segment sequence code; S4: Calling the path segment sequence code to calculate the fitting degree of the vehicle path sequence record collected by the vehicle GPS, screening vehicle resources of the same sequence and binding the corresponding vehicle delivery path label; S5: Based on the replenishment rhythm window matching judgment, a delivery instruction is sent, the consistency between the delivery path and the vehicle delivery path label is monitored in real time, and a transportation scheduling adjustment result is generated according to the monitoring result.

2. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: The node congestion score includes the vehicle transportation time, the route congestion index, and the node load compensation value; the replenishment rhythm window includes the rhythm window start time, the rhythm window end time, and the rhythm window alternative time period; the path segment sequence code is specifically the picking sort code, the channel occupancy sequence number, and the shipping time sequence number; the vehicle delivery path label includes the path sequence fitting degree, the vehicle matching sequence number, and the path binding identifier; the transportation scheduling adjustment result includes the distribution path matching degree, the vehicle path consistency, and the real-time scheduling deviation.

3. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: The specific steps of real-time monitoring of the transportation time of logistics vehicles, inputting the vehicle transportation time into the principal component analysis model, calculating the congestion score of the transportation route, supplementing the node load score data, and generating the congestion score include: S101: monitoring the start and end times of a transport task of a logistics vehicle, and analyzing the duration of a single vehicle operation of the transport task based on the deviation between the start and end times of the transport task obtained by monitoring, to obtain the transport duration; S102: Based on the transport duration input into the principal component analysis model, deviation comparison is performed in combination with a preset duration threshold range of the corresponding transport route to establish a route congestion analysis result; S103: performing point mapping processing on the nodes associated with the transport route according to the route congestion analysis result, and generating a congestion point of the node's current transport route.

4. The logistics and transportation management method based on multi-source data fusion according to claim 3 is characterized in that: The deviation comparison is performed on the preset time threshold range of the transportation route using the formula: Calculate the congestion level I of the node's current route and obtain the congestion level analysis result of the node's current route; Where n is the total number of sample vehicles, i is the index of the vehicle, and U T =μ T +2σ T is the upper threshold, L T =μ T -2σ T is the lower threshold, μ T is the average transport time of transport vehicles, μ is the standard sign of the mean, σ is the standard deviation, δ is the penalty coefficient, ε is the reward coefficient, μ+2σ is the upper threshold, μ-2σ is the lower threshold, max(T i -U T ,0) means that if the transportation time exceeds the upper threshold, the corresponding reward is calculated, max(L T -T i ,0) means if the transport time is lower than the lower threshold, the corresponding penalty is calculated, C adj represents the average congestion index of neighboring nodes, C base Expressed as the baseline congestion index.

5. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: Obtain the estimated arrival start and end times of replenishment tasks within the next transportation cycle, determine overlapping replenishment time periods based on the congestion score of the transportation route, and generate multiple candidate rhythm window time periods based on the start and end times. The specific steps for determining the replenishment rhythm window include: S201: Arrival Time: Obtain the estimated arrival time and departure time corresponding to each replenishment task in the next transportation cycle, classify the start and end times of each replenishment task according to the transportation cycle number, and locate the time boundaries of the start and end time intervals of the replenishment tasks in each cycle to generate the replenishment time interval; S202: Based on the obtained replenishment time interval, the node congestion score is called to determine whether each replenishment time interval coincides with the score of the corresponding node, and replenishment tasks with overlapping time periods are filtered and aggregated to obtain conflicting time intervals. S203: Based on the conflicting time interval, the replenishment time interval is called, and time axis shift processing is performed on the boundaries of both ends of the conflicting time interval, including adjusting the starting boundary time and the ending boundary time of the conflicting time interval as a shift operation, to generate multiple non-overlapping new time intervals, and eliminating windows that intersect with the start and end times of the replenishment task to obtain rhythm candidate intervals; S204: Based on the obtained rhythm candidate intervals, all candidate intervals are arranged in ascending order according to the start time, and time window combinations that do not overlap within a single cycle are screened to generate a replenishment rhythm window.

6. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: The specific steps of matching the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse end shipping schedule are as follows: the warehouse end picking and delivery sequence, the transportation channel usage sequence, and the shipping schedule data are serialized and sorted to generate the path segment sequence code. S301: Obtain the replenishment estimated arrival start time and the replenishment estimated arrival end time in the replenishment rhythm window, match the start and end times of the replenishment tasks in the replenishment rhythm window with the scheduled shipping time of the warehouse end, call the warehouse task node information to perform logical judgment on the matching group, group the task combinations with overlapping time into the same rhythm group, and generate a time matching result between the replenishment tasks and the warehouse nodes; S302: Based on the obtained time matching result between the replenishment task and the warehouse node, the warehouse picking plan number, transportation channel code, and shipping appointment time corresponding to the task in the rhythm group are called, and the three types of data are bound to perform task association sequence. The parameters in the task group are jointly sorted according to the time sequence relationship to obtain the operation rhythm sorting sequence of the associated nodes after sorting; S303: According to the generated operation rhythm sorting sequence, the warehouse number, channel number and shipping number in the sorting sequence are called, and the path segment structure information is generated in sequence. The path fields in the same rhythm group are marked with unique picking sorting codes, and the sequence segment number is established and the corresponding task path mapping relationship is recorded to generate the path segment sequence code.

7. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: The specific steps of calling the path segment sequence code to perform similarity calculation on the vehicle path sequence record collected by the vehicle GPS, screening vehicle resources with similar sequences and binding the corresponding vehicle to release the path label include: S401: Calling the path segment sequence code, obtaining the vehicle path sequence record collected by the vehicle-mounted GPS, comparing the matching degree of the three types of fields in the path segment, namely, the numbered node, the time sequence, and the channel number, calculating the structural coincidence rate between the current path segment and the preset reference value of the vehicle path segment, and generating the path fit; S402: Based on the acquired path fit, the idle time and available loading status parameters of each vehicle in the vehicle status list in the current cycle are retrieved, and vehicles with path fit higher than the path matching threshold and meeting the status conditions are selected as a candidate set to establish vehicle matching resources; S403: Based on the obtained vehicle matching resources, the path matching threshold is determined by calculating the distribution characteristics through the time difference data, and the unique identification code of each candidate vehicle and the corresponding path segment sequence code are called to generate a label combination. The combined labels are grouped by vehicle and written into the vehicle task allocation field to establish a vehicle deployment path label.

8. The logistics and transportation management method based on multi-source data fusion according to claim 7 is characterized in that: The current path segment and the preset reference value of the vehicle path segment are processed using the formula: Calculate the structural overlap rate S overlap ; Among them, N match It is the node number set corresponding to the real-time data of the vehicle's current path status and the preset benchmark of the vehicle path segment. The number of node number matches formed by counting the number of identical numbers is T. max is the time difference tolerance threshold, which is obtained by statistical quantiles, N total Indicates the total number of nodes in the current path segment. represents the mean of the time sequence difference, C match Indicates the number of channel number matches, C total Indicates the total number of channels in the current path segment.

9. The logistics and transportation management method based on multi-source data fusion according to claim 1 is characterized in that: Based on the replenishment rhythm window matching judgment, sending a delivery instruction, monitoring the consistency of the delivery path and the vehicle delivery path label in real time, and generating a transportation scheduling adjustment result according to the monitoring result include: S501: Obtain the arrival start and end times of the replenishment task in the replenishment rhythm window, call the vehicle delivery path tag, match the start and end times of the replenishment task with the path segment times recorded in the corresponding vehicle path tag, and measure the time difference to generate a rhythm matching offset result for the current path task; S502: Based on the obtained rhythm matching offset result, the route task numbers with time offsets less than the offset threshold are screened, the corresponding vehicle numbers and replenishment task numbers are called to execute the combinations, a delivery instruction is generated for each combination, the route number and issuance time are recorded, and a delivery task is established; S503: Calling the obtained delivery task, collecting the current positioning trajectory of the delivery vehicle and synchronously comparing it with the preset path segment sequence number, extracting the current vehicle operation number of each path segment and the delivery tag path segment number to perform node consistency judgment, obtaining the ratio of the number of consistent nodes in each path segment to the total number, and generating a path consistency ratio; S504: Based on the obtained path consistency ratio, the path segment sequence code is called, the path segment number is replaced based on the real-time scheduling deviation of the path deviation benchmark ratio, the path segment offset result is replaced and recorded and appended to the control scheduling log to generate the transportation scheduling adjustment result.

10. A logistics and transportation management system based on multi-source data fusion, characterized in that: The system is used to implement the logistics and transportation management method of multi-source data fusion according to any one of claims 1 to 9, and the system includes: The node load integration module monitors the transport time of logistics vehicles in real time, inputs the vehicle transport time into the principal component analysis model, calculates the congestion score of the transport route, supplements the node load integration data, and transmits the generated congestion score to the replenishment rhythm window module; The replenishment rhythm window module obtains the estimated replenishment arrival start and end times of the replenishment task in the next transportation cycle, determines the overlapping replenishment time periods based on the congestion score of the transportation route, performs shifting based on the start and end times to generate multiple rhythm window candidate time periods, and passes them to the path segment sequence generation module; The path segment sequence generation module matches the replenishment estimated arrival start time and replenishment estimated arrival end time in the replenishment rhythm window with the warehouse-side shipping scheduled time, and serializes and sorts the warehouse-side picking and outbound order, the transportation channel usage order, and the shipping scheduled time data to generate a path segment sequence code and pass it to the vehicle label binding module; The vehicle label binding module calls the current path segment sequence in the path segment sequence code, compares the fit with the path segment sequence of the previous period, selects vehicle resources with similar sequences, binds the corresponding vehicle delivery path labels, and transmits them to the collaborative scheduling control module; The collaborative scheduling control module sends delivery instructions based on the replenishment rhythm window matching judgment, monitors the consistency of the delivery path and the vehicle delivery path label in real time, and generates transportation scheduling adjustment results based on the monitoring results.

Images