Logistics-to-delivery time prediction system and method based on multi-data fusion

Abstract

The application discloses a logistics-to-delivery time prediction system and method based on multi-data fusion, relates to the technical field of logistics-to-delivery time prediction, and aims to solve the technical problem that the prediction result of an existing logistics-to-delivery time prediction system is greatly deviated from the actual delivery time. A data acquisition module, a path planning module, a time calculation module and a warning and output module are arranged. The output end of the data acquisition module is connected with the path planning module and the time calculation module respectively and is used for acquiring multi-source data. The multi-source data includes historical logistics data, order and delivery information data, weather data, real-time road condition data, map data, transportation tool characteristic data, cargo special demand data, transportation tool connection implicit loss data and transportation tool coordination delay correlation data. The output end of the path planning module is connected with the time calculation module. The application has the advantage of improving the accuracy of delivery time prediction in a multi-modal transport scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics delivery time prediction, more particularly, to a logistics delivery time prediction system and method based on multi-data fusion. BACKGROUND

[0002] In the modern logistics industry, especially in the multimodal transport scenario, accurate prediction of delivery time is of great significance to improve logistics efficiency, optimize customer experience and ensure the stability of the supply chain. However, the existing logistics delivery time prediction technology has significant shortcomings in the multimodal transport scenario: due to the connection and cooperation of multiple transportation tools (such as trucks, airplanes, ships, etc.) in multimodal transport, the existing technology often fails to systematically quantify the hidden losses in the connection process of transportation tools (such as waiting time caused by differences in operation processes at different nodes, information transmission delay, etc.), and lacks effective correlation analysis and integrated calculation of transportation tool coordination delays (such as the cascading effect of previous airplane delays on subsequent trucks), resulting in a large deviation between the prediction results and the actual delivery time, making it difficult to meet the precise logistics scheduling and customer's fine-grained demand for delivery time. In view of this, we propose a logistics delivery time prediction system and method based on multi-data fusion. SUMMARY

[0003] The purpose of the present application is to provide a logistics delivery time prediction system and method based on multi-data fusion to solve the technical problem of large deviation between the prediction results of the existing logistics delivery time prediction system and the actual delivery time.

[0004] To solve the above technical problems, the present application provides the following technical scheme: a logistics delivery time prediction system based on multi-data fusion, comprising: a data acquisition module, a path planning module, a time calculation module, and a warning and output module;

[0005] The output end of the data acquisition module is connected with the path planning module and the time calculation module respectively, for acquiring multi-source data, the multi-source data including historical logistics data, order and shipping information data, weather data, real-time traffic data, map data, transportation tool characteristic data, cargo special demand data, transportation tool connection hidden loss data, and transportation tool coordination delay correlation data;

[0006] The output end of the path planning module is connected with the time calculation module, for generating and optimizing multimodal transport candidate paths based on the information of the data acquisition module, including connection node planning and hidden loss evaluation of different transportation tools;

[0007] The output end of the time calculation module is connected with a pre-warning and output module, for fusing multi-source data to calculate the estimated delivery time of each candidate path, including differential correction based on the characteristics of the transportation tools, transfer time adjustment combining the special needs of the goods, quantitative connection with implicit loss, and collaborative efficiency fluctuation correction based on the collaborative delay correlation model;

[0008] The pre-warning and output module is used for generating pre-warning information and final delivery time according to the estimated delivery time, and pushing to the customer terminal and the logistics management terminal.

[0009] Preferably, the data acquisition module comprises:

[0010] A historical data acquisition sub-module is used for acquiring historical logistics data, the historical logistics data comprising transportation paths, mileages and historical time consumptions of each section, transportation tool types and characteristic parameters, transfer node stay time, transportation tool connection implicit loss historical data, and transportation tool collaborative delay data;

[0011] The transportation paths cover complete route trajectories of trunk transportation and branch distribution; the mileages and historical time consumptions of each section contain time consumption differences of different transportation tools in the same section; the transportation tool types and characteristic parameters cover driving speed limits, load capacities, and environmental adaptation ranges of different tools; the transfer node stay time distinguishes time consumption differences of ordinary goods and special goods in the storage and loading and unloading links; the transportation tool connection implicit loss historical data includes implicit time consumption records of waiting time and information transmission delay caused by operation process differences of different nodes; and the transportation tool collaborative delay data covers influence rules of the delay of a preceding tool on a subsequent tool in different transportation tool combinations.

[0012] An order and shipment information acquisition sub-module is used for acquiring current order and shipment information, the shipment information comprising goods information, transportation tool combination schemes, and connection node operation requirements;

[0013] The goods information contains physical attributes, storage requirements, and transportation restrictions of the goods; the transportation tool combination schemes specify the types and connection sequences of transportation tools in each section; and the connection node operation requirements specify document processing procedures, loading and unloading operation standards, and personnel qualification conditions in detail; the sub-module comprises an influence factor extraction unit, which is used for identifying and extracting influence factors affecting the delivery time from the acquired information.

[0014] Preferably, the data acquisition module further comprises:

[0015] A real-time data acquisition sub-module is used for acquiring real-time road condition data, the real-time road condition data comprising changes in the passing state and passing capacity of each section, real-time weather data, transfer node real-time throughput data, transportation tool real-time state data, connection node real-time implicit loss data, and pre-transportation tool delay pre-warning data.

[0016] Among them, real-time weather data covers the weather phenomenon and change trend of the area along the transportation route; the real-time throughput data of the transfer node includes the current cargo handling capacity and equipment operation status of each node; the real-time state data of the transportation tool includes location information, running state and abnormal situation feedback; the real-time implicit loss data of the connection node covers the operation efficiency of the current operation personnel, the real-time delay situation of the document delivery; the delay warning data of the previous transportation tool includes the delay reason and the expected recovery time;

[0017] The map data acquisition submodule is used to acquire geographical map data of the starting point and the ending point, and the geographical map data includes road network, airport / port distribution and operation capacity, geographical coordinates of multimodal transport connection node and transfer process complexity level, etc.

[0018] Among them, the road network includes the connection relationship and traffic conditions of different grade roads; the airport / port distribution and operation capacity cover the throughput, operation period and equipment configuration of the hub; the geographical coordinates of the multimodal transport connection node specify the specific location and surrounding traffic connection of the node; the transfer process complexity level is divided according to the factors such as security check link and loading / unloading equipment configuration, and the transfer operation difficulty of different nodes is specified.

[0019] Preferably, the historical data acquisition submodule further includes a data screening unit and a data updating unit:

[0020] The data screening unit is used to screen target historical data from historical logistics data;

[0021] In the screening process, firstly, based on the transportation tool combination scheme of the current order and the connection node type, similar cases are preliminarily matched from the historical data;

[0022] Then, the connection node type similarity is calculated to evaluate the consistency of the historical node and the current node in the operation process and equipment configuration;

[0023] At the same time, the transportation tool coordination combination similarity is calculated to consider the matching degree of the historical transportation tool combination and the current scheme in the connection logic and time efficiency characteristics;

[0024] Finally, the historical data meeting the matching degree requirement is selected as the target data by comprehensively considering various similarity indexes;

[0025] The data updating unit is used to update the historical data according to the preset period, and in the updating process, firstly, new transportation case data in the period is collected, including time consumption record, implicit loss situation and coordination delay instance of each link;

[0026] Then, the efficiency data of the new document circulation system is included in the implicit loss historical data, and the latest parameters of the coordination delay correlation model are supplemented.

[0027] Meanwhile, the historical data exceeding the storage period or not matching the current transportation standard is cleaned up to ensure the timeliness and applicability of the historical database.

[0028] Preferably, the influence factor extracted by the influence factor extraction unit includes:

[0029] Implicit loss factor of connection node: personnel efficiency loss rate during the night period, i.e., the degree of change of personnel operation efficiency during night operation relative to daytime;

[0030] Docking friction coefficient of cross-enterprise transport parties, used to measure the coordination efficiency of different enterprises in the aspects of document delivery and responsibility division;

[0031] Coordination fluctuation sensitive factor: the subsequent truck waiting threshold when the previous tool is an airplane, i.e., when the airplane delay time reaches the threshold, the subsequent truck needs to adjust the departure plan, and the threshold is determined based on the critical time length of historical airplane delay leading to truck adjustment;

[0032] Railway scheduling elasticity coefficient when the ship is late, used to evaluate the scheduling adjustment capability of railway transportation under the condition of ship delay and the influence degree on the arrival time.

[0033] Preferably, the path planning module includes:

[0034] Multi-modal combined analysis unit, used to generate an adapted transportation tool combination scheme, in the analysis process, first, according to the type of goods, transportation distance and urgency, the applicable transportation tool type is preliminarily screened, then combined with the timeliness characteristics, cost factors and connection feasibility of each transportation tool, a plurality of combination schemes are constructed, and finally through evaluating the comprehensive adaptability of each scheme, a plurality of potential transportation tool combinations are selected;

[0035] Candidate path generation unit, used to generate a plurality of candidate logistics paths, in generating the paths, first, based on the geographical positions of the starting point and the ending point, combined with the road network and hub distribution in the map data, the driving routes of each transportation tool are planned, then the transfer nodes of each segment of transportation tool are determined, the transfer process and operation steps at the nodes are determined, and the implicit loss risk points in the path are identified, including the connection nodes during personnel shift period, the transfer hubs during equipment maintenance period, and finally the complete candidate path including each segment mileage, connection node, transfer process and risk point is formed;

[0036] The path optimization subunit is configured to optimize the candidate paths, wherein in the optimization process, firstly, a calculation method of the connection implicit loss index is determined, the overall implicit loss level is evaluated based on the implicit loss proportion and the occurrence frequency of each connection node of the path in the historical data, the collaborative fluctuation resistance is calculated, the influence range and degree of the preceding tool delay on the subsequent transportation link in the path are considered, then the corresponding weight is assigned to each optimization index, the score of each path is calculated comprehensively, and finally, the high-score paths of a preset number are selected as the core candidate paths according to the score ranking.

[0037] Preferably, the time calculation module comprises:

[0038] The transportation tool differentiation correction unit is configured to perform characteristic correction on the segments of different transportation tools in the core candidate paths, for the truck segment, the basic time consumption of the segment is adjusted according to the road grade, the real-time road condition and the driving characteristics of the truck, for the airplane segment, the flight time consumption is corrected in combination with the air route weather condition, the air traffic control condition and the flight characteristics of the airplane, and for the ship segment, the navigation and berthing time consumption is adjusted according to the channel condition, the port operation state and the navigation characteristics of the ship.

[0039] The goods special demand processing unit is configured to calculate the differentiated residence time of the transfer node, for the cold-chain goods, the residence duration at the node is determined according to the storage temperature requirement, the preservation period and the temperature control equipment state of the transfer node, and for the dangerous goods, the residence time is calculated according to the safety management regulation, the inspection process and the state of the special processing facility of the node.

[0040] The implicit loss quantification unit is configured to quantify the implicit time consumption of the connection link, wherein in the calculation process, firstly, the type of the document flow is determined according to the operation specification of the connection node, and the document flow efficiency coefficient is applied, then the proficiency level of the operation personnel is determined in combination with the real-time personnel configuration information, the proficiency correction coefficient is applied, and the transfer process complexity correction value is calculated according to the transfer process complexity level, wherein the basic time consumption of the multi-level security check is increased by a corresponding proportion compared with the simple connection, and finally, the implicit time consumption of the connection link is obtained by comprehensively considering the coefficients and the correction values.

[0041] The collaborative fluctuation correction unit is configured to calculate the interlocking time efficiency influence between the transportation tools, for the airplane delay and the delay duration greater than the subsequent truck waiting threshold, the truck waiting time is calculated, otherwise, the subsequent transportation is performed according to the original plan, and for the ship delay, the subsequent railway transportation time consumption is adjusted in combination with the railway dispatching elasticity coefficient.

[0042] The full-link integration unit is used to aggregate the time consumption after each segment correction, that is, the driving time of each transportation tool segment after characteristic correction, superimpose the differentiated residence time of the transfer node, that is, the node residence time calculated according to the type of goods, add the implicit loss time consumption of the connection link, that is, the implicit time consumption calculated by the quantization unit, count the collaborative fluctuation correction value, that is, the waiting time, the railway adjustment time consumption or the scheme switching compensation value obtained by the collaborative fluctuation correction unit, and finally integrate all time elements to form the preliminary estimated arrival time.

[0043] Preferably, the time calculation module further comprises a dynamic correction unit for correcting the preliminary estimated arrival time in combination with real-time data. When real-time implicit loss changes occur, such as document transmission delay caused by sudden system failure, the corresponding time change value is calculated according to the influence range and duration of the failure. When collaborative delay upgrade occurs, such as the expansion of the previous tool delay time causing the subsequent tool to miss the scheduled shift, the alternative shift and dispatch adjustment time of the subsequent tool are analyzed to determine the collaborative delay upgrade correction value. The dynamic correction unit superimposes the above change value and correction value to the preliminary estimated arrival time to obtain the corrected arrival time.

[0044] Preferably, the early warning and output module comprises:

[0045] The multi-dimensional early warning unit is used to generate special early warnings for different transportation tools and goods types.

[0046] When the connection implicit loss early warning is triggered, that is, the document circulation delay of a certain node exceeds the set threshold, the system first analyzes the delay reason. If it is caused by the type of document, the system pushes the suggestion to switch the electronic document first, and simultaneously attaches the expected time consumption improvement after switching;

[0047] When the collaborative delay early warning is triggered, that is, the delay of the previous tool may cause a chain reaction in the subsequent link, the system immediately retrieves alternative subsequent transportation tools and paths to generate a time efficiency comparison table of alternative schemes, and clearly shows the expected arrival time and adjustment cost of each scheme.

[0048] The output unit is used to select the optimal result from the corrected arrival time. During the selection process, the time efficiency stability, implicit loss proportion and collaborative fluctuation influence degree of each path are comprehensively considered, and the arrival time with the best overall performance is selected. The generated detailed time composition report includes the time consumption proportion of each transportation tool segment, the specific composition of the implicit loss of the connection link, the influence time and reason analysis of the collaborative fluctuation on the overall time, and the report is pushed to the customer terminal in synchronization, facilitating the customer to view in real time. At the same time, it is pushed to the logistics management terminal to provide decision reference for dispatch personnel.

[0049] A logistics arrival time prediction method based on multi-data fusion, comprising the following steps:

[0050] S1, collect multi-source data, the multi-source data including historical logistics data, order and delivery information data, weather data, real-time traffic data, map data, transportation tool characteristic data, goods special demand data, transportation tool connection implicit loss data, transportation tool coordination delay correlation data;

[0051] S2, generate and optimize multi-modal transport candidate paths based on the collected multi-source data, including planning the connection nodes of different transportation tools and evaluating the implicit loss;

[0052] S3, fuse multi-source data to calculate the estimated delivery time of each candidate path, including differential correction based on transportation tool characteristics, transit time adjustment combined with goods special requirements, quantification of connection implicit loss, and coordination efficiency fluctuation correction based on a coordination delay correlation model;

[0053] S4, generate warning information and final delivery time according to the estimated delivery time, and push the warning information and final delivery time to the customer terminal and the logistics management terminal.

[0054] Compared with the prior art, the beneficial effects of the present application are:

[0055] 1. The present application quantifies the influence of transportation tool connection implicit loss and coordination delay by fusing historical logistics data, real-time traffic, weather data and other multi-source information, significantly improves the accuracy of delivery time prediction in multi-modal transport scenarios, and effectively solves the problem of large prediction deviation in the prior art.

[0056] 2. Based on the accurate delivery time prediction result, the present application optimizes the multi-modal transport path planning, preferentially selects paths with low implicit loss and strong coordination fluctuation resistance, reduces unnecessary transportation delay and resource waste, and further improves the overall efficiency of logistics transportation.

[0057] 3. The present application dynamically corrects the estimated time in combination with real-time data and generates multi-dimensional warnings, which can respond to sudden situations in the transportation process in time, improve the flexibility of logistics scheduling, and at the same time push detailed time composition reports to the customer terminal to enhance the customer's perception and trust of the delivery time. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 The figure is a schematic diagram of the system framework of the present application. DETAILED DESCRIPTION

[0059] In order to facilitate those skilled in the art to understand the technical scheme of the present application, the technical scheme of the present application will be further described in conjunction with the drawings of the specification.

[0060] Example 1, as Figure 1As shown, the present application provides a logistics-to-delivery time prediction system based on multi-data fusion, comprising a data acquisition module, a path planning module, a time calculation module, and a warning and output module.

[0061] The output end of the data acquisition module is connected with the path planning module and the time calculation module, respectively, for collecting multi-source data, including historical logistics data, order and delivery information data, weather data, real-time traffic data, map data, transportation tool characteristic data, cargo special requirement data, transportation tool connection implicit loss data, and transportation tool coordination delay correlation data.

[0062] The output end of the path planning module is connected with the time calculation module, for generating and optimizing multi-modal transport candidate paths based on the information of the data acquisition module, including connection node planning and implicit loss evaluation of different transportation tools.

[0063] The output end of the time calculation module is connected with the warning and output module, for calculating the estimated delivery time of each candidate path by fusing multi-source data, including differential correction based on transportation tool characteristics, transit time adjustment combined with cargo special requirements, connection implicit loss quantification, and coordination efficiency fluctuation correction based on the coordination delay correlation model.

[0064] The warning and output module is used to generate warning information and final delivery time according to the estimated delivery time, and push it to the customer terminal and the logistics management terminal.

[0065] In the embodiment of the present application, the data acquisition module comprises:

[0066] A historical data acquisition sub-module is used to collect historical logistics data, including transportation paths, mileage and historical time consumption of each section, transportation tool types and characteristic parameters, transit node stay time, transportation tool connection implicit loss historical data, and transportation tool coordination delay data.

[0067] Among them, the transportation path covers the complete route trajectory of trunk transportation and branch distribution; the mileage and historical time consumption of each section contain the time consumption difference of different transportation tools on the same section; the transportation tool type and characteristic parameter cover the driving speed limit, carrying capacity and environmental adaptability range of different tools; the transit node stay time distinguishes the time consumption difference of ordinary goods and special goods in the warehouse and loading and unloading links; the transportation tool connection implicit loss historical data includes implicit time consumption records such as waiting time and information transmission delay caused by operation process differences of different nodes; and the transportation tool coordination delay data covers the influence law of the delay of the previous tool on the subsequent tool in different transportation tool combinations.

[0068] An order and delivery information collection sub-module is configured to collect current order and delivery information, and the delivery information includes cargo information, a transport tool combination scheme, and a connection node operation requirement;

[0069] The cargo information includes physical attributes, storage requirements, and transportation restrictions of the cargo, the transport tool combination scheme specifies the type and connection sequence of each transport tool, and the connection node operation requirement specifies a document processing flow, loading and unloading operation standards, and personnel qualification conditions in detail.

[0070] A real-time data collection sub-module is configured to collect real-time road condition data, and the real-time road condition data includes the passing state and passing capacity change of each road section, real-time weather data, real-time throughput data of a transfer node, real-time state data of a transport tool, real-time implicit loss data of a connection node, and pre-transport tool delay warning data.

[0071] The real-time weather data covers weather phenomena and change trends in the areas along the transport route, the real-time throughput data of the transfer node includes the current cargo processing capacity and equipment operation state of each node, the real-time state data of the transport tool includes location information, running state, and abnormal situation feedback, and the real-time implicit loss data of the connection node covers the operation efficiency of the current operation personnel and the real-time delay of document delivery.

[0072] A map data collection sub-module is configured to obtain geographical map data of the starting point and the ending point, and the geographical map data includes a road network, airport / port distribution and operation capacity, geographical coordinates of a multimodal transport connection node, and a transfer process complexity level.

[0073] The road network includes the connection relationship and passing conditions of different grade roads, the airport / port distribution and operation capacity covers the throughput, operation period, and equipment configuration of the hub, the geographical coordinates of the multimodal transport connection node specify the specific location and surrounding traffic connection of the node, and the transfer process complexity level is divided according to factors such as security check links and loading and unloading equipment configuration, and specifies the transfer operation difficulty of different nodes.

[0074] In the embodiment of the present application, the historical data collection sub-module further includes a data screening unit and a data updating unit.

[0075] The data screening unit is configured to screen target historical data from historical logistics data.

[0076] In the screening process, first, similar cases are preliminarily matched from the historical data based on the transport tool combination scheme and the connection node type of the current order.

[0077] Then, the consistency of the historical nodes and the current nodes in the work process and the device configuration is evaluated by calculating the similarity of the connection node types.

[0078] Meanwhile, the similarity of the transportation tool combination is calculated to measure the matching degree of the historical transportation tool combination and the current scheme in the connection logic and the timeliness characteristics.

[0079] Finally, the historical data that meets the matching degree requirement is selected as the target data by comprehensively considering the similarity indexes.

[0080] The similarity calculation formula is as follows:

[0081] ;

[0082] wherein, represents the comprehensive similarity, which is the overall quantitative evaluation result of the matching degree of the historical logistics data and the current order data, and the value range is , and the value closer to 1 indicates a higher matching degree;

[0083] represents the similarity of the connection node types, which is used to measure the consistency of the connection nodes in the historical logistics data and the connection nodes involved in the current order in the work process specification, the device configuration standard, the personnel operation process, etc., and the value range is ;

[0084] represents the similarity of the transportation tool combination, which is used to evaluate the similarity of the transportation tool combination scheme in the historical logistics data and the transportation tool combination adopted by the current order in the connection sequence, the tool type matching, and the collaborative work logic, and the value range is ;

[0085] represents the weight coefficient of the similarity of the connection node types, which is used to reflect the importance of the connection node types in the overall matching evaluation, and is set according to the influence of the node factors on the timeliness in different transportation scenarios, and the value range is ;

[0086] represents the weight coefficient of the similarity of the transportation tool combination, which is used to reflect the importance of the transportation tool combination scheme in the overall matching evaluation, and is set according to the influence of the tool cooperation on the timeliness fluctuation, and the value range is , and satisfies ;

[0087] The data updating unit is used to update the historical data at a preset period. During the updating process, the new transportation case data in the period is collected, including the time consumption records, the implicit loss conditions, and the collaborative delay instances in each link.

[0088] Then, the efficiency data of the new document circulation system is incorporated into the implicit loss historical data, and the latest parameters of the collaborative delay correlation model are supplemented;

[0089] At the same time, historical data that exceeds the retention period or does not match the current transportation standard is cleaned up to ensure the timeliness and applicability of the historical database.

[0090] In an embodiment of the present application, the influence factor extracted by the influence factor extraction unit includes:

[0091] The interface node implicit loss factor: the personnel efficiency loss rate during the night period, that is, the degree of change of personnel operation efficiency during night work relative to daytime;

[0092] The cross-enterprise transportation party interface friction coefficient, used to measure the collaborative efficiency of different enterprises in document transfer, responsibility division, etc.

[0093] The calculation formula is:

[0094] ;

[0095] Wherein, represents the night personnel efficiency loss rate, which is used to quantify the difference between personnel operation efficiency during night work and daytime operation. When the value is greater than 1, it means that the night efficiency is lower than the daytime, and the larger the value, the more serious the efficiency loss;

[0096] represents the average operation time during the night, which refers to the average time for completing the standard operation process of the interface node during the night period (usually 22:00 of the current day to 6:00 of the next day), and the unit is minute;

[0097] represents the average operation time during the day, which refers to the average time for completing the same standard operation process of the interface node during the daytime period (usually 6:00 to 22:00 of the current day), and the unit is minute;

[0098] ;

[0099] Wherein, represents the cross-enterprise interface friction coefficient, which is used to measure the efficiency loss degree of collaborative operation between different enterprises. When the value is greater than 1, it means that the cross-enterprise collaborative time is higher than that of the same enterprise, and the larger the value, the more significant the friction loss;

[0100] represents the average cross-enterprise collaborative time, which refers to the average time for different enterprises to perform collaborative operations such as document transfer, responsibility confirmation, and cargo transfer at the interface node, and the unit is minute;

[0101] Indicates the average time-consuming of the same cooperation in the same enterprise, that is, the average time-consuming of the same cooperation in the same enterprise at the connection node, and the unit is minute;

[0102] The cooperative fluctuation sensitive factor is a subsequent truck waiting threshold when the previous tool is an airplane, that is, when the airplane delay time reaches the threshold, the subsequent truck needs to adjust the departure plan, and the threshold is determined based on the critical time length of the historical airplane delay leading to the truck adjustment;

[0103] The railway scheduling elasticity coefficient when the ship is late, which is used to evaluate the scheduling adjustment capability of the railway transportation under the condition of the ship being late and the influence degree on the arrival time;

[0104] The calculation formula is:

[0105]

[0106] Among them, The truck waiting threshold of the airplane delay, that is, when the airplane delay time reaches the value, the subsequent truck needs to start the departure plan adjustment process, and the unit is minute;

[0107] The historical airplane delay time mean indicates the arithmetic mean of the actual delay time of the airplane at the same connection node in the past period, and the unit is minute;

[0108] The historical delay time standard deviation is used to measure the dispersion degree of the historical airplane delay time, and the unit is minute; 1.28 is the critical value coefficient of one-sided 90% confidence interval, which ensures that the threshold covers most of the delay scenarios that need to be adjusted;

[0109]

[0110] Among them, The railway scheduling elasticity coefficient is used to evaluate the scheduling adjustment capability of the railway transportation when the ship is late, and the value is greater than 1, which indicates that the time-consuming increases after adjustment, and the smaller the value is, the better the elasticity is and the higher the adjustment efficiency is;

[0111] The railway adjustment time-consuming after the ship is late indicates the total time-consuming of the scheduling plan adjustment, resource reconfiguration and other operations of the railway transportation for cooperation after the ship is late, and the unit is minute;

[0112] The normal railway scheduling time-consuming indicates the standard time-consuming of the railway transportation to complete the scheduling preparation according to the original plan under the condition that the ship is on time, and the unit is minute.

[0113] In the embodiment of the application, the path planning module comprises:

[0114] ​​The multi-modal transport combination analysis unit is used to generate an adaptive transport tool combination scheme. In the analysis process, first, the applicable transport tool type is preliminarily screened according to the cargo type, transport distance and urgency, and then a plurality of combination schemes are constructed by combining the time efficiency characteristics, cost factors and connection feasibility of each transport tool. Finally, a plurality of potential transport tool combinations are selected by evaluating the comprehensive adaptability of each scheme.

[0115] The candidate path generation unit is used to generate a plurality of candidate logistics paths. In generating the paths, first, the driving routes of each transport tool are planned based on the geographic locations of the starting point and the ending point, combined with the road network and hub distribution in the map data. Then, the transfer nodes of each segment of the transport tool are determined, the transfer process and operation steps at the nodes are determined, and the implicit loss risk points in the path are identified, including the transfer nodes at the personnel shift time period and the transfer hubs at the equipment maintenance time period. Finally, a complete candidate path including the mileage of each segment, the transfer nodes, the transfer process and the risk points is formed.

[0116] The path optimization sub-unit is used to optimize the candidate paths. In the optimization process, first, the calculation method of the connection implicit loss index is determined. Based on the implicit loss proportion and occurrence frequency of each connection node in the path in the historical data, the overall implicit loss level is evaluated. Then, the collaborative fluctuation resistance is calculated to consider the influence range and degree of the delay of the previous tool in the path on the subsequent transport link. Then, each optimization index is assigned a corresponding weight, and the score of each path is calculated comprehensively. Finally, according to the score, a plurality of high-score paths are selected as the core candidate paths.

[0117] The comprehensive score formula is:

[0118] ;

[0119] Wherein, represents the path comprehensive score, which is a quantitative evaluation result of the overall advantages and disadvantages of the candidate path, and the value range is [0, 100]. The higher the score, the better the comprehensive performance of the path in implicit loss control and collaborative fluctuation resistance.

[0120] represents the connection implicit loss index, which is used to measure the overall level of implicit loss of each connection node in the path, and the value range is [0, 100]. The lower the value, the smaller the implicit loss. The calculation method is the weighted average value of the implicit loss values of each node (the weight is set according to the importance of the node).

[0121] represents the collaborative fluctuation resistance, which is used to evaluate the resistance of the path to the collaborative delay of the transport tool, and the value range is [0, 100]. The higher the value, the stronger the resistance. It is calculated based on the delay transmission probability matrix in the collaborative delay correlation model (the lower the transmission probability, the higher the resistance).

[0122] a represents a weight coefficient of the connection implicit loss index, is used to reflect the importance of the implicit loss factor in the path optimization, is set according to the proportion of the influence of the implicit loss on the total time in the transportation scene, and the value range is ;

[0123] a represents a weight coefficient of the cooperative fluctuation resistance, is used to reflect the importance of the cooperative fluctuation factor in the path optimization, is set according to the influence degree of the cooperative delay on the total time, and the value range is , and satisfies .

[0124] In the embodiment of the application, the time calculation module comprises:

[0125] The transportation tool differentiation correction unit is used for performing characteristic correction on the segments of different transportation tools in the core candidate path, for the truck segment, adjusting the basic time consumption of the segment according to the road grade, real-time road condition and driving characteristics of the truck, for the airplane segment, correcting the flight time consumption in combination with the weather condition of the route, air traffic control condition and navigation characteristics of the airplane, and for the ship segment, adjusting the navigation and berthing time consumption according to the channel condition, port operation state and navigation characteristics of the ship;

[0126] The goods special demand processing unit is used for calculating the differentiated residence time of the transfer node, for the cold chain goods, determining the residence time at the node according to the storage temperature requirement, preservation period and temperature control equipment state of the transfer node, and for the dangerous goods, calculating the residence time according to the safety management regulation, inspection process and special processing facility state of the node;

[0127] The implicit loss quantification unit is used for quantifying the implicit time consumption of the connection link, in the calculation process, first, the type of the document flow (electronic or paper) is determined according to the operation specification of the connection node, and the document flow efficiency coefficient is applied, then the proficiency level of the operation personnel is determined in combination with the real-time personnel configuration information, and the proficiency correction coefficient is applied, and at the same time, the transfer process complexity correction value is calculated according to the transfer process complexity level, wherein the basic time consumption of the multiple-level security check is increased by a corresponding proportion compared with the simple connection, and finally, the implicit time consumption of the connection link is obtained by comprehensively considering various coefficients and correction values;

[0128] The calculation formula is:

[0129] ;

[0130] Wherein, a represents the total implicit time consumption of the connection link, is the sum of all implicit losses in the connection link, the unit is minute, and reflects the non-explicit time consumption caused by the factors such as document flow, personnel operation and process complexity.

[0131] represents the single document flow efficiency coefficient, used to distinguish the flow efficiency difference between electronic single and paper single, electronic single takes , paper single takes The smaller the value, the higher the flow efficiency;

[0132] represents the basic connection time consumption, refers to the benchmark time for completing the core operation of the connection node under standard conditions (daytime operation, cooperation with the same enterprise, electronic single), unit is minute, based on historical data statistics;

[0133] represents the transfer process complexity correction value, used to quantify the time consumption difference caused by different transfer process complexity levels, unit is minute, multi-level security check scene usually increases 50%-150% of the basic time consumption of simple connection scene;

[0134] The coordination fluctuation correction unit is used to calculate the chain time efficiency influence between transportation tools. When a certain transportation tool is delayed, the corresponding correction mechanism needs to be started according to the type of transportation tool and the length of delay. If it is an airplane delay and the length of delay , the waiting time of the truck is calculated, otherwise the original plan is executed. For the late ship, the subsequent railway transportation time consumption is adjusted in combination with the railway scheduling flexibility coefficient;

[0135] The calculation formula is:

[0136] ;

[0137] Among them, represents the subsequent truck waiting time, refers to the additional time consumption of the subsequent truck waiting for connection due to the delay of the airplane, unit is minute, if the delay of the airplane does not reach the threshold, the waiting time is 0;

[0138] represents the actual delay time of the airplane, refers to the difference between the actual arrival time and the planned arrival time of the airplane, unit is minute, positive value represents delay, negative value represents advance;

[0139] represents the coordination influence coefficient, taken from the parameters of the coordination delay correlation model, used to quantify the influence degree of airplane delay on the subsequent truck waiting time, the value range is , such as 0.7 represents that the truck waiting time is 70% of the airplane delay time;

[0140] ;

[0141] Among them, Railway adjustment time after ship delay, which refers to the total time of railway transportation rescheduling caused by ship delay, in minutes, reflecting the response cost of railway transportation to ship delay;

[0142] Original railway plan time, which refers to the standard time of railway transportation preparation according to the original plan under the condition of ship on-time, in minutes, based on historical scheduling data;

[0143] If there is an alternative for the subsequent tool (such as replacing a delayed airplane with a high-speed train), the time efficiency compensation value of the synchronous switching scheme is calculated:

[0144] ;

[0145] Where, Time efficiency compensation value of scheme switching, which refers to the additional total time when using alternative transportation tools, in minutes, used to evaluate the time efficiency cost of scheme adjustment;

[0146] Resource allocation time, which refers to the time required for resource coordination, transportation capacity arrangement, etc. for scheduling alternative transportation tools, in minutes, directly related to the availability of alternative tools;

[0147] Transfer connection time, which refers to the operation time such as loading and unloading, document change, etc. during the transfer from the original transportation tool to the alternative tool, in minutes, affected by the type of goods and the efficiency of the transfer node;

[0148] The whole link integration unit is used to integrate the time after each segment correction, i.e. the driving time of each transportation tool segment after characteristic correction, plus the differentiated residence time of the transfer node, i.e. the node residence time calculated according to the type of goods, plus the implicit loss time of the connection link, i.e. the implicit time calculated by the quantization unit, plus the cooperative fluctuation correction value, i.e. the waiting time, railway adjustment time or scheme switching compensation value obtained by the cooperative fluctuation correction unit, finally integrating all time elements to form the preliminary estimated arrival time;

[0149] The calculation formula is:

[0150] ;

[0151] Where, Preliminary estimated arrival time, which refers to the total time of the whole link estimated arrival after integrating all time elements, in minutes, is the basis value for subsequent dynamic correction;

[0152] The sum of the time after correction of each transportation tool segment, which refers to the sum of the driving time of each segment of truck, airplane, ship, etc. after characteristic correction, in minutes;

[0153] represents the sum of differentiated residence time of each node, refers to the sum of residence time calculated by all transit nodes according to the special requirements of goods, the unit is minute, and the special goods such as cold chain and dangerous goods are distinguished from ordinary goods;

[0154] represents the waiting time of the truck, and the subsequent truck waiting time calculated by the cooperative fluctuation correction unit is referred to, the unit is minute.

[0155] In the embodiment of the application, the time calculation module further comprises a dynamic correction unit for correcting the preliminary estimated arrival time in combination with real-time data, when real-time implicit loss changes occur, such as document transmission delay caused by sudden document system failure, according to the influence range and duration of the failure, the corresponding time change value is calculated, when cooperative delay upgrade occurs, such as the expansion of the previous tool delay time causing the subsequent tool to miss the original scheduled shift, the replacement shift and dispatching adjustment time of the subsequent tool are analyzed to determine the cooperative delay upgrade correction value, and the dynamic correction unit superimposes the above change value and correction value to the preliminary estimated arrival time to obtain the corrected arrival time;

[0156] The calculation formula is:

[0157]

[0158] Among them, represents the corrected arrival time, which is the final estimated total arrival time after dynamic adjustment by real-time data, the unit is minute, and the influence of initial estimation and various sudden factors is comprehensively reflected;

[0159] represents the real-time implicit loss change value, which refers to the additional time consumption caused by sudden implicit loss events (such as document system failure, temporary shortage of personnel, etc.), the unit is minute, and the value is only taken when the sudden situation occurs, and is 0 when there is no change;

[0160] represents the cooperative delay upgrade correction value, which refers to the additional adjustment time consumption when the delay time of the current sequence tool is further expanded and causes the subsequent tool to be delayed in a chain (such as missing the original shift), the unit is minute, and its calculation refers to the influence coefficient of delay upgrade in the cooperative delay correlation model, and is 0 when there is no upgrade.

[0161] In the embodiment of the application, the early warning and output module comprises:

[0162] The multi-dimensional early warning unit is used for generating special early warning for different transport tools and goods types;

[0163] ​When the implicit loss warning is triggered, that is, the single document flow delay of a node exceeds the set threshold, the system first analyzes the delay reason. If it is caused by the type of single document, the system pushes the suggestion of switching to electronic single document in priority, and simultaneously attaches the expected time consumption improvement after switching;

[0164] When the collaborative delay warning is triggered, that is, the previous tool delay may cause a chain reaction of the subsequent link, the system immediately retrieves the alternative subsequent transport tools and paths, generates a time efficiency comparison table of the alternative schemes, and clearly shows the expected arrival time and adjustment cost of each scheme;

[0165] The warning threshold calculation formula is:

[0166] ;

[0167] Among them, represents the warning threshold, which is the time critical value for determining whether to trigger the warning, and the unit is minute. When the actual estimated arrival time or the real-time monitored time consumption exceeds the value, the system starts the warning mechanism;

[0168] represents the historical maximum deviation rate, which refers to the maximum deviation percentage of the actual arrival time and the preliminary estimated arrival time in the same transportation scene within a certain period, and the value range is (such as 0.2 represents 20%), which reflects the extreme situation of time efficiency fluctuation in historical data, and is used to ensure that the warning threshold covers high-risk scenarios;

[0169] The output unit is used to filter the optimal result from the corrected arrival time. In the filtering process, the time efficiency stability, implicit loss proportion and collaborative fluctuation influence degree of each path are comprehensively considered, and the arrival time with the best comprehensive performance is selected to generate a detailed time composition report. The report includes the time consumption proportion of each transport tool segment, the specific composition of implicit loss in the connection link, the influence time and reason analysis of collaborative fluctuation on the overall time, and is pushed to the customer terminal and the logistics management terminal at the same time, so that the customer can view the report in real time, and the dispatch personnel can make decision reference.

[0170] In the embodiment of the application, the self-optimization function of the data updating unit includes:

[0171] After each actual delivery is completed, the historical database of implicit loss in the connection is automatically updated, the actual implicit loss data of each connection node in this delivery is collected, including single document flow time consumption, personnel operation efficiency, etc. The new data is compared with the historical data, the difference reason is analyzed, and the coefficients of single document flow, personnel configuration and other factors are adjusted;

[0172] The coefficient optimization formula is:

[0173] ;

[0174] wherein, represents the optimized coefficient, refers to the new coefficient value after calibration by the actual delivery data, used to improve the accuracy of subsequent implicit loss calculation, covering the single document circulation efficiency coefficient, personnel proficiency correction coefficient;

[0175] represents the pre-optimization coefficient, refers to the original coefficient value used by the system before this delivery, which is the benchmark data for optimization calculation;

[0176] represents the predicted implicit time consumption, refers to the estimated implicit loss time calculated by the system based on the pre-optimization coefficient before this delivery, in minutes, which is the benchmark for comparison with the actual implicit time consumption;

[0177] represents the predicted implicit time consumption, refers to the estimated implicit loss time calculated by the system based on the pre-optimization coefficient before this delivery, in minutes, which is the benchmark for comparison with the actual implicit time consumption;

[0178] Based on actual collaborative delay cases, iteratively optimize the collaborative influence coefficient, the truck waiting threshold of aircraft delay, and the railway dispatching elasticity coefficient. Collect the specific circumstances of each collaborative delay, including the delay duration, the scope of influence, and the treatment results. By analyzing the correlation between the combination of transportation tools and collaborative influence in the cases, adjust the relevant coefficients and thresholds. At the same time, according to the actual time consumption of scheme switching, optimize the calculation logic of the time efficiency compensation model. Through continuous iteration, improve the prediction accuracy of the system for the chain reaction between transportation tools, and reduce the prediction error caused by collaborative delay.

[0179] Embodiment 2, a logistics delivery time prediction method based on multi-data fusion, comprising the following steps:

[0180] S1, collect multi-source data, the multi-source data includes historical logistics data, order and delivery information data, weather data, real-time traffic data, map data, transportation tool characteristics data, cargo special demand data, transportation tool implicit loss data, transportation tool collaborative delay correlation data;

[0181] S2, generate and optimize the multi-modal transport candidate path based on the collected multi-source data, including planning the connection nodes of different transportation tools and evaluating the implicit loss;

[0182] S3, fuse multi-source data to calculate the estimated delivery time of each candidate path, including differential correction based on transportation tool characteristics, transit time adjustment combined with cargo special requirements, quantification of implicit loss, and collaborative efficiency fluctuation correction based on collaborative delay correlation model;

[0183] S4, generating early warning information and final arrival time according to the estimated arrival time, and pushing the early warning information and final arrival time to the customer terminal and the logistics management terminal.

[0184] Initial source, derivation and logical relationship of each formula

[0185] I. Comprehensive similarity calculation formula

[0186] Formula content:

[0187] ;

[0188] Initial source: derived from the weighted summation model in mathematics, which is a basic method for multi-index comprehensive evaluation, widely used in similarity matching, decision analysis and other scenarios.

[0189] Derivation and modification: this application adapts it to the logistics scene, and defines:

[0190] To connect node type similarity (measure the consistency of node operation process and equipment configuration);

[0191] For transport tool collaborative combination similarity (measure the matching degree of tool combination connection logic and time efficiency characteristics);

[0192] Weight According to the influence degree of node factors and tool cooperation on time efficiency, precise matching of historical data and current order is realized.

[0193] II. Night personnel efficiency loss rate calculation formula

[0194] Formula content:

[0195] ;

[0196] Initial source: derived from the basic ratio calculation of efficiency comparison, which belongs to a simple proportion relationship model in mathematics, used to quantify the relative difference between two similar indicators.

[0197] Derivation and modification: this application is used for personnel efficiency evaluation of logistics connection nodes.

[0198] Defined as the average time of standard operation at night (22:00-6:00 the next day);

[0199] Defined as the average time of the same operation during the day (6:00-22:00);

[0200] Through the ratio Directly reflect the loss degree of night efficiency relative to daytime (the ratio is less than 1, indicating that the efficiency of night operation is lower than that of daytime operation) Nighttime efficiency is lower.

[0201] Three, the friction coefficient calculation formula of cross-enterprise docking

[0202] Formula content:

[0203] ;

[0204] Initial source: Same as "Nighttime Personnel Efficiency Loss Rate", derived from the proportional relationship model, used to measure the efficiency difference between different subjects.

[0205] Derivation change: This application focuses on inter-enterprise collaboration efficiency.

[0206] Defined as the average time consumption of cross-enterprise collaborative operations (document transfer, responsibility division, etc.);

[0207] Defined as the average time consumption of the same operation within the enterprise;

[0208] Through the ratio Quantify the friction loss of cross-enterprise collaboration ( Indicates that the cross-enterprise collaboration time is higher).

[0209] Four, the formula for calculating the truck waiting threshold of aircraft delay

[0210] Formula content:

[0211] ;

[0212] Initial source:

[0213] Derived from the confidence interval calculation of normal distribution in statistics, specifically the upper limit formula of one-sided 90% confidence interval ( , where is the critical value).

[0214] Derivation change: This application is used to determine the critical value of truck plan adjustment after aircraft delay.

[0215] is the historical aircraft delay duration mean;

[0216] is the historical delay duration standard deviation;

[0217] is the critical value coefficient of one-sided 90% confidence interval, ensuring that the threshold covers most delay scenarios that need to be adjusted, making become a quantitative standard for triggering truck plan adjustment.

[0218] Five, the formula for calculating the elasticity coefficient of railway dispatching

[0219] Formula content:

[0220] ;

[0221] Initial source: derived from the proportional relationship model, used to measure the relative difference between the adjusted time consumption and the benchmark time consumption.

[0222] Derivation change: this application is used to evaluate the scheduling adjustment capability of the railway when the ship is late.

[0223] Defined as the total time consumption of railway adjustment after the ship is late;

[0224] Defined as the normal railway scheduling time consumption when the ship is on time;

[0225] Ratio Directly reflects the scheduling flexibility The smaller, the better the flexibility, the higher the adjustment efficiency.

[0226] Sixth, the calculation formula of the comprehensive score of the path

[0227] Formula content:

[0228] ;

[0229] Initial source: derived from the weighted comprehensive evaluation model, which is a commonly used linear weighted summation method in multi-index decision-making.

[0230] Derivation change:

[0231] This application is used for the optimization of candidate paths:

[0232] For the convergence implicit loss index (the lower the value, the better);

[0233] For the collaborative fluctuation resistance (the higher the value, the better);

[0234] And Respectively, the weight of the two (according to the influence degree on the total time efficiency), through the comprehensive score Sort and screen the optimal path.

[0235] Seven, the calculation formula of the total time consumption of convergence implicit

[0236] Formula content:

[0237] ;

[0238] Initial source: derived from the time consumption calculation model of multiple factors superposition, based on the product relationship of the basic value and the correction coefficient.

[0239] Derivation of changes:

[0240] This application is used to quantify the implicit loss of the link:

[0241] The efficiency coefficient of the document circulation (difference between electronic and paper documents);

[0242] (Night efficiency loss rate) and (Cross-enterprise friction coefficient) are defined impact factors;

[0243] The basic connection time under standard conditions;

[0244] The transfer process complexity correction value (such as the increased time consumption of multi-level security check);

[0245] Based on the above factors, the total implicit time consumption of the connection link is directly calculated .

[0246] Eight, the formula for calculating the waiting time of the truck caused by the delay of the airplane

[0247] Formula content:

[0248] ;

[0249] Initial source: linear relationship model under conditional judgment, combined with threshold condition and proportional coefficient calculation results.

[0250] Derivation of changes:

[0251] This application is used to calculate the impact of airplane delay on subsequent trucks:

[0252] The actual delay time of the airplane;

[0253] The waiting threshold defined in the previous section;

[0254] The coordination influence coefficient ( ), quantifying the transmission ratio of airplane delay to truck waiting;

[0255] Determine whether to calculate the waiting time through conditional judgment, realize the quantitative conduction of coordinated delay.

[0256] Nine, the formula for calculating the adjustment time consumption of the railway after the delay of the ship

[0257] Formula content:

[0258] ;

[0259] Initial source: derived from the product model of the base value and the correction coefficient, used to proportionally adjust the benchmark value.

[0260] Derivation change:

[0261] This application is used to calculate the impact of ship delay on the railway:

[0262] The original railway plan time for the on-time ship;

[0263] The railway scheduling flexibility coefficient defined in the foregoing;

[0264] The adjusted railway time is calculated directly by multiplication , reflecting the impact of scheduling flexibility on time efficiency.

[0265] Ten, scheme switching time compensation value calculation formula

[0266] Formula content:

[0267] ;

[0268] Initial source: derived from the addition model, used to summarize the total time of multiple independent links.

[0269] Derivation change:

[0270] This application is used to evaluate the additional cost of transportation tool alternatives:

[0271] The resource allocation time (the time to coordinate the alternative tool);

[0272] The transfer connection time (loading and unloading, document change, etc.);

[0273] The sum is the total time compensation value of scheme switching, used for decision reference of alternative scheme.

[0274] Eleven, preliminary estimated arrival time calculation formula

[0275] Formula content:

[0276]

[0277] Initial source: derived from the total link time accumulation model, the total time is obtained by summarizing the time of each link.

[0278] Derivation change: this application is used to calculate the preliminary estimated time of the whole path:

[0279] The total time of each transportation tool segment after correction;

[0280] Sum of differentiated dwell time for each node

[0281] Total hidden consumption time for connection

[0282] The value in the bracket is the collaborative fluctuation correction value (truck waiting, railway adjustment, scheme switching compensation, etc.)

[0283] Sum That is, the initial estimated delivery time of the whole link.

[0284] Twelfth, the formula for calculating the corrected delivery time

[0285] Formula content:

[0286] ;

[0287] Initial source: derived from the dynamic adjustment model, superimposed on the basis of initial value real-time variation.

[0288] Derivation change: this application is used to optimize the estimated time combined with real-time data:

[0289] The preliminary estimated delivery time

[0290] Real-time hidden consumption change value (such as delay caused by bill system failure)

[0291] Collaborative delay upgrade correction value (such as additional adjustment time caused by the expansion of previous delay)

[0292] Finally, the dynamically corrected delivery time is obtained .

[0293] Thirteen, the formula for calculating the warning threshold

[0294] Formula content:

[0295] ;

[0296] Initial source: derived from the safety margin calculation model, superimposed on the basis of the historical maximum deviation ratio.

[0297] Derivation change: this application is used to set the warning trigger condition:

[0298] The preliminary estimated delivery time

[0299] The historical maximum deviation rate (reflecting the historical extreme fluctuation)

[0300] Threshold value Ensure coverage of high-risk scenarios, trigger early warning when actual time consumption approaches or exceeds this value.

[0301] Fourteenth, coefficient optimization formula

[0302] Formula content:

[0303] ;

[0304] Initial source: derived from feedback correction model, adjust the coefficient by the ratio of actual value and predicted value.

[0305] Derivation changes: this application is used for system self-optimization:

[0306] For the coefficient before optimization (such as the efficiency coefficient of document circulation);

[0307] For the actual implicit time consumption, For the predicted time consumption based on ;

[0308] Through proportional correction, get , so that the coefficient continuously adapts to the actual scene and improves the subsequent calculation accuracy.

[0309] Logical relationship between formulas

[0310] Basic factor layer: formulas two, three, four and five are the core influencing factors, which quantify the basic characteristics such as personnel efficiency, enterprise cooperation, delay threshold and scheduling flexibility, and are the input parameters for subsequent calculation.

[0311] Path and time consumption calculation layer: formula six is used for path optimization; formulas seven, eight, nine and ten calculate the specific time consumption of implicit loss and cooperative delay, which provide basis for whole link time calculation.

[0312] Total time calculation layer: formula eleven integrates the time consumption of each link to get the preliminary estimated time; formula twelve gets the final time through dynamic correction; formula thirteen sets the early warning threshold to form a complete time efficiency prediction and early warning logic. Self-optimization layer: formula fourteen uses actual delivery data to optimize the basic factors in reverse, so that the formula parameters of the whole system are continuously iterated, and the prediction accuracy is improved.

[0313] Each formula forms a closed-loop logic of "basic factor → path selection → time consumption calculation → dynamic correction → early warning → self-optimization", which jointly supports the logistics-to-delivery time prediction system based on multi-data fusion.

[0314]

[0315] ​The embodiments of the present application disclose the preferred embodiments, but are not limited to the same. Those skilled in the art can easily understand the spirit of the present application according to the above embodiments, and make different inferences and changes, as long as they do not deviate from the spirit of the present application, and are within the protection scope of the present application.

Claims

1. A logistics-to-delivery time prediction system based on multi-data fusion, characterized by, The application relates to a multi-modal transport time estimation system, comprising: a data acquisition module, a path planning module, a time calculation module, and a warning and output module; the output end of the data acquisition module is connected with the path planning module and the time calculation module, and is used for acquiring multi-source data, wherein the multi-source data comprises historical logistics data, order and delivery information data, weather data, real-time road condition data, map data, transport tool characteristic data, goods special demand data, transport tool connection implicit loss data, and transport tool cooperation delay correlation data; the output end of the path planning module is connected with the time calculation module, and is used for generating and optimizing multi-modal transport candidate paths based on the information of the data acquisition module, wherein the candidate paths comprise connection node planning and implicit loss evaluation of different transport tools; the output end of the time calculation module is connected with the warning and output module, and is used for fusing the multi-source data to calculate estimated delivery times of the candidate paths, wherein the estimated delivery times comprise differential correction based on transport tool characteristics, transit time adjustment combined with goods special demands, implicit loss quantification, and cooperation efficiency fluctuation correction, and the cooperation efficiency fluctuation correction is based on a cooperation delay correlation model; the warning and output module is used for generating warning information and final delivery times according to the estimated delivery times, and pushing the warning information and the final delivery times to customer terminals and logistics management terminals; the time calculation module comprises: a transport tool differential correction unit, which is used for performing characteristic correction on segments of different transport tools in a core candidate path, adjusting the basic time consumption of a truck segment according to road grades, real-time road conditions and truck driving characteristics, correcting flight time consumption of an airplane segment combined with air route weather conditions, air traffic control conditions and airplane navigation characteristics, and adjusting navigation and berthing time consumption of a ship segment according to channel conditions, port operation states and ship navigation characteristics; a goods special demand processing unit, which is used for calculating differential residence times of transit nodes, determining the residence time of a cold-chain goods node according to storage temperature requirements, preservation time limits and the temperature control equipment state of the transit node, and calculating the residence time of a dangerous goods node according to safety management regulations, inspection procedures and the state of special processing facilities of the node; an implicit loss quantification unit, which is used for quantifying implicit time consumption of a connection link, wherein, in the calculation process, the type of a document flow is determined according to the operation specification of a connection node, a document flow efficiency coefficient is applied, the proficiency level of operation personnel is determined combined with real-time personnel configuration information, a proficiency correction coefficient is applied, and a transfer process complexity correction value is calculated according to the complexity level of a transfer process, wherein a multi-level security check is added with a corresponding proportion of basic time consumption compared with simple connection, and finally, implicit time consumption of the connection link is obtained by comprehensively applying various coefficients and correction values; a cooperation fluctuation correction unit, which is used for calculating the interlocking time efficiency influence between transport tools, wherein, when an airplane is delayed and the delay time is longer than a subsequent truck waiting threshold, truck waiting time is calculated, otherwise, the truck is executed according to the original plan, and when a ship is delayed, the subsequent railway transport time consumption is adjusted combined with a railway dispatching elasticity coefficient. The full-link integration unit is used for integrating the time consumption after each segment correction, i.e. the driving time of each transportation tool segment after characteristic correction, superimposing the differentiated residence time of the transfer node, i.e. the node residence time length calculated according to the cargo type, adding the implicit loss time consumption of the connection link, i.e. the implicit time consumption calculated by the quantization unit, adding the cooperative fluctuation correction value, i.e. the waiting time length, the railway adjustment time consumption or the scheme switching compensation value obtained by the cooperative fluctuation correction unit, and finally integrating all time elements to form the preliminary estimated delivery time.

2. The logistics-to-delivery time prediction system based on multi-data fusion according to claim 1, characterized in that, The data acquisition module comprises: a historical data acquisition sub-module, used for acquiring historical logistics data, wherein the historical logistics data comprises a transportation path, mileages and historical time consumption of each road segment, a transportation tool type and characteristic parameters, transfer node residence time, transportation tool connection implicit loss historical data and transportation tool cooperative delay data; wherein the transportation path covers the complete route track of trunk transportation and branch distribution; the mileages and historical time consumption of each road segment include the time consumption difference of different transportation tools in the same road segment; the transportation tool type and characteristic parameters cover the driving speed limit, load capacity and environmental adaptation range of different tools; the transfer node residence time distinguishes the time consumption difference of ordinary goods and special goods in the storage and loading and unloading links; the transportation tool connection implicit loss historical data includes the implicit time consumption records of the waiting time caused by the operation process difference of different nodes and the information transmission delay; and the transportation tool cooperative delay data covers the influence law of the delay of the previous tool on the subsequent tool in different transportation tool combinations; an order and delivery information acquisition sub-module, used for acquiring current order and delivery information, wherein the delivery information comprises cargo information, a transportation tool combination scheme and connection node operation requirements; wherein the cargo information includes the physical properties, storage requirements and transportation restrictions of the goods; the transportation tool combination scheme specifies the type and connection order of each transportation tool; and the connection node operation requirements specifically stipulate the document processing flow, loading and unloading operation standards and personnel qualification conditions; the sub-module comprises an influence factor extraction unit, used for identifying and extracting the influence factors affecting the delivery time from the acquired information.

3. The logistics-to-delivery time prediction system based on multi-data fusion according to claim 2, characterized in that, The data acquisition module further comprises: a real-time data acquisition sub-module, used for acquiring real-time road condition data, wherein the real-time road condition data comprises the traffic state and traffic capacity change of each road segment, real-time weather data, transfer node real-time throughput data, transportation tool real-time state data, connection node real-time implicit loss data and previous transportation tool delay warning data; wherein the real-time weather data covers the weather phenomena and change trend of the areas along the transportation route; the transfer node real-time throughput data includes the current cargo processing capacity and equipment operating state of each node; the transportation tool real-time state data includes location information, operating state and abnormal situation feedback; the connection node real-time implicit loss data covers the operation efficiency of the current operation personnel and the real-time delay situation of the document transmission; and the previous transportation tool delay warning data includes the delay reason and predicted recovery time; The map data collection submodule is configured to obtain geographical map data of the starting point and the ending point, and the geographical map data includes a road network, an airport / port distribution and operation capacity, geographical coordinates of a multimodal transport connection node, and a transfer process complexity level, etc. The road network includes connection relationships and traffic conditions of different levels of roads; the airport / port distribution and operation capacity covers the throughput, operation period, and equipment configuration of the hub; the geographical coordinates of the multimodal transport connection node specify the specific location and surrounding traffic connection of the node; and the transfer process complexity level is divided according to factors such as security check links and loading / unloading equipment configuration, and specifies the difficulty of transfer operation at different nodes.

4. The logistics-to-delivery time prediction system based on multi-data fusion according to claim 3, characterized in that, The historical data collection submodule further includes a data screening unit and a data updating unit. The data screening unit is configured to screen target historical data from historical logistics data. In the screening process, firstly, similar cases are preliminarily matched from the historical data based on the transport tool combination scheme of the current order and the connection node type; Then, the consistency of the historical node and the current node in the operation process and the equipment configuration is evaluated by calculating the connection node type similarity; Meanwhile, the matching degree of the historical transport tool combination and the current scheme in the connection logic and time efficiency characteristics is considered by calculating the transport tool combination similarity; Finally, the historical data that meets the matching degree requirement is selected as the target data by comprehensively considering various similarity indexes; The data updating unit is configured to update the historical data at a preset period, and in the updating process, firstly, new transport case data in the period is collected, including the time consumption record, hidden loss condition, and collaborative delay instance of each link; Then, the efficiency data of the new document flow system is included in the hidden loss historical data, and the latest parameters of the collaborative delay correlation model are supplemented; Meanwhile, the historical data that exceeds the storage period or does not match the current transport standard is cleaned up to ensure the timeliness and applicability of the historical database.

5. The logistics-to-delivery time prediction system based on multi-data fusion according to claim 2, characterized in that, The influence factors extracted by the influence factor extraction unit include: A connection node hidden loss factor: a personnel efficiency loss rate in a night period, that is, the change degree of personnel operation efficiency in the night period relative to the day period; A cross-enterprise transport party interface friction coefficient for measuring the collaborative efficiency of different enterprises in the document transfer and responsibility division links; A collaborative fluctuation sensitive factor: a subsequent truck waiting threshold when the previous tool is an airplane, that is, when the airplane delay time reaches the threshold, the subsequent truck needs to adjust the departure plan, and the threshold is determined based on the critical time length of the historical airplane delay leading to the truck adjustment; A ship delay railway dispatching elasticity coefficient for evaluating the dispatching adjustment ability of railway transportation under the ship delay condition and the influence degree on the arrival time.

6. The logistics-to-delivery time prediction system based on multi-data fusion of claim 1, wherein, The path planning module includes: A multimodal transport combination analysis unit configured to generate an adaptive transport tool combination scheme, and in the analysis process, firstly, the applicable transport tool type is preliminarily screened according to the cargo type, transport distance, and emergency level, then a plurality of combination schemes are constructed by combining the time efficiency characteristics, cost factors, and connection feasibility of each transport tool, and finally, a plurality of potential transport tool combinations are selected by evaluating the comprehensive adaptability of each scheme. The candidate path generating unit is configured to generate a plurality of candidate logistics paths, wherein, when generating the paths, firstly, based on geographical positions of the starting point and the ending point, in combination with a road network and hub distribution in map data, travel routes of various transport tools are planned, then, connection nodes of the various transport tools are determined, transfer processes and operation steps at the nodes are determined, and implicit loss risk points in the paths are identified, including connection nodes in personnel shift periods and transfer hubs in equipment maintenance periods, and finally, complete candidate paths including various path segments, connection nodes, transfer processes and risk points are formed; The path optimization subunit is configured to optimize the candidate paths, wherein, in the optimization process, firstly, a calculation method of a connection implicit loss index is determined, based on implicit loss proportions and occurrence frequencies of the connection nodes in the paths in historical data, a whole implicit loss level is evaluated, then, a collaborative fluctuation resistance is calculated, and an influence range and degree of a previous tool delay on a subsequent transport link in the paths are considered, then, various optimization indexes are assigned with corresponding weights, scores of each path are calculated comprehensively, and finally, according to the scores, a preset number of high-score paths are selected as core candidate paths.

7. The logistics-to-delivery time prediction system based on multi-data fusion according to claim 6, characterized in that, The time calculation module further includes a dynamic correction unit configured to correct the preliminary estimated delivery time in combination with real-time data, when real-time implicit loss changes occur, such as a sudden bill system failure leading to bill transmission delay, according to an influence range and duration of the failure, a corresponding time change value is calculated, when collaborative delay escalation occurs, such as a previous tool delay time expansion leading to a subsequent tool missing a scheduled shift, a replacement shift and dispatch adjustment time consumption of the subsequent tool are analyzed, a collaborative delay escalation correction value is determined, and the dynamic correction unit superimposes the above change value and correction value to the preliminary estimated delivery time to obtain a corrected delivery time.

8. The logistics-to-delivery time prediction system based on multi-data fusion of claim 1, wherein, The warning and output module includes: A multi-dimensional warning unit configured to generate special warnings for different transport tools and cargo types; When a connection implicit loss warning is triggered, that is, a node bill transfer delay exceeds a set threshold, the system firstly analyzes the delay reason, if the delay is caused by the bill type, a suggestion of preferentially switching to an electronic bill is pushed, and an improved time consumption after switching is attached; When a collaborative delay warning is triggered, that is, a previous tool delay may cause a chain reaction of subsequent links, the system immediately searches for alternative subsequent transport tools and paths, generates a time efficiency comparison table of the alternative schemes, and clearly shows the estimated delivery time and adjustment cost of each scheme; An output unit configured to select an optimal result from the corrected delivery time, wherein, in the selection process, time efficiency stability, implicit loss proportion and collaborative fluctuation influence degree of each path are considered comprehensively, an optimal delivery time is selected, a detailed time composition report is generated, and the report includes time consumption proportions of various transport tool segments, specific compositions of connection link implicit losses, influence time and reason analysis of collaborative fluctuation on the whole time, the report is pushed to a client terminal in real time for the client to view, and is also pushed to a logistics management terminal for providing decision reference for dispatch personnel.

9. A method of predicting the time of arrival of a shipment based on a multi-data fusion-based logistics-to-delivery time prediction system as claimed in claim 1, characterized in that, The method includes the following steps: S1, collect multi-source data, the multi-source data including historical logistics data, order and delivery information data, weather data, real-time traffic data, map data, transportation tool characteristic data, goods special demand data, transportation tool connection implicit loss data, transportation tool coordination delay correlation data; S2, generate and optimize multi-modal transport candidate paths based on the collected multi-source data, including planning the connection nodes of different transportation tools and evaluating the implicit loss; S3, fuse the multi-source data to calculate the estimated delivery time of each candidate path, including differentiating correction based on transportation tool characteristics, adjusting transit time based on goods special demand, quantifying connection implicit loss, and correcting coordination efficiency fluctuation based on a coordination delay correlation model; S4, generate early warning information and final delivery time according to the estimated delivery time, and push the early warning information and final delivery time to the customer terminal and logistics management terminal.

Images