A cargo volume prediction method and device, electronic equipment and readable storage medium

Abstract

This disclosure relates to the field of logistics and transportation technology, and proposes a freight volume prediction method, device, electronic device, and readable storage medium. The method includes: acquiring historical waybill data, and generating a multi-dimensional initial probability transition matrix based on the historical waybill data, the initial probability transition matrix including historical route resource information and route delivery probability information; acquiring real-time waybill barcode scanner operation data, and determining the planned route resource information, real-time vehicle status, and real-time freight volume summary data for the current day based on the real-time waybill barcode scanner operation data; adjusting the route delivery probability information of the initial probability transition matrix based on the planned route resource information for the current day, historical route resource information, and real-time vehicle status to obtain a target probability transition matrix; and determining the expected freight volume allocated to each target route based on the real-time freight volume summary data and the target probability transition matrix. The technical solutions provided by one or more embodiments of this disclosure can accurately predict the allocated freight volume of each route.

Description

Technical Field

[0001] This disclosure relates to the field of logistics and transportation technology, specifically to a cargo volume forecasting method, apparatus, electronic device, and readable storage medium. Background Technology

[0002] With rapid economic development, people's demand for express delivery is increasing, and competition in the logistics industry is gradually intensifying. Cost reduction and efficiency improvement have become an industry consensus. For logistics companies, facing freight demand, it is necessary to choose a more reasonable transportation solution among different modes of transportation, taking into account time requirements and transportation costs.

[0003] In the logistics and transportation sector, labor costs, vehicle costs, and site costs are significant components of production and operating costs. Therefore, optimizing human resources, efficiently utilizing vehicle resources, and rationally planning transportation routes can have an immediate effect on reducing costs and increasing efficiency in the logistics industry.

[0004] Against this backdrop, accurate control over site cargo volume data has become increasingly important, especially obtaining cargo volume data several hours in advance. Therefore, accurate cargo volume forecasting data has become a very urgent need. Summary of the Invention

[0005] In view of this, one or more embodiments of the present disclosure provide a freight volume prediction method, apparatus, electronic device and readable storage medium that can accurately predict the allocated freight volume of each route, thereby optimizing the resource allocation of vehicles and sites.

[0006] This disclosure provides a freight volume forecasting method, comprising: acquiring historical waybill data for a predetermined time period, and generating a multi-dimensional initial probability transition matrix based on the historical waybill data, the initial probability transition matrix including historical route resource information and route delivery probability information; acquiring real-time waybill barcode scanner operation data, and determining planned route resource information, real-time vehicle status, and real-time freight volume summary data for the current day based on the real-time waybill barcode scanner operation data; adjusting the route delivery probability information of the initial probability transition matrix based on the planned route resource information for the current day, the historical route resource information, and the real-time vehicle status to obtain a target probability transition matrix; and determining the expected freight volume allocation for each target route based on the real-time freight volume summary data and the target probability transition matrix.

[0007] This disclosure also provides a cargo volume prediction device, comprising: a matrix building unit, configured to acquire historical waybill data for a predetermined time period, and generate a multi-dimensional initial probability transition matrix based on the historical waybill data, the initial probability transition matrix including historical route resource information and route shipment probability information; a data acquisition unit, configured to acquire real-time waybill barcode scanner operation data, and determine the planned route resource information, real-time vehicle status, and real-time cargo volume summary data for the current day based on the real-time waybill barcode scanner operation data; a matrix updating unit, configured to adjust the route shipment probability information of the initial probability transition matrix based on the planned route resource information for the current day, the historical route resource information, and the real-time vehicle status, to obtain a target probability transition matrix; and a cargo volume prediction unit, configured to determine the expected cargo volume allocated to each target route based on the real-time cargo volume summary data and the target probability transition matrix.

[0008] This disclosure also provides an electronic device including a memory and a processor, the memory for storing a computer program, which, when executed by the processor, implements the above-described cargo volume prediction method.

[0009] This disclosure also provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the above-described cargo volume prediction method.

[0010] The technical solutions provided by one or more embodiments of this disclosure can acquire a large amount of historical waybill data, establish a multi-dimensional initial probability transition matrix, and ensure the reliability of the prediction method. Through real-time waybill barcode scanner operation data, the route resource information, vehicle status information, and freight volume summary information for the day can be determined. Based on this, and using the planned route resource information for the day, historical route resource information, and real-time vehicle status, the probability distribution of the initial probability transition matrix can be optimized and adjusted to obtain the target probability transition matrix, ensuring the real-time performance and accuracy of the prediction method. Using real-time freight volume summary data and the target probability transition matrix, the expected freight volume allocated to each target route can be accurately predicted, facilitating the determination of the expected arrival freight volume at each site, and thus optimizing the resource allocation of vehicles and sites.

[0011] The technical solutions provided by one or more embodiments of this disclosure can accurately predict the expected freight volume allocated to each target route, thereby improving the utilization efficiency of route resources. Based on the predicted freight volume data, vehicle resources can be dynamically and more rationally allocated, making the addition and removal of vehicles more precise and efficient. The expected freight volume allocated to each target route also helps to determine the expected arrival volume of freight at each site, which can reduce the operating costs of the sites.

[0012] The technical solutions provided by one or more embodiments of this disclosure, based on predicted cargo volume data, can output full load and low load alarms in advance, allowing sufficient time for the site to handle abnormal situations, thereby reducing the risk of cargo falling off the site and improving customer satisfaction.

[0013] The technical solutions provided by one or more embodiments of this disclosure can be connected to operations planning algorithms based on real-time predicted cargo volume data at the route dimension, and can perform fast and slow loading in advance according to the predicted cargo volume, thereby reducing the difficulty of downstream unloading operations. Attached Figure Description

[0014] The features and advantages of the embodiments of this disclosure will be more clearly understood by referring to the accompanying drawings, which are illustrative and should not be construed as limiting the present disclosure in any way. In the drawings:

[0015] Figure 1 A schematic diagram illustrating the steps of a cargo volume forecasting method in one embodiment of this disclosure is shown.

[0016] Figure 2 A flowchart illustrating a cargo volume forecasting method according to one embodiment of this disclosure is shown;

[0017] Figure 3 A schematic diagram of the functional modules of a cargo volume prediction device in one embodiment of this disclosure is shown;

[0018] Figure 4 A schematic diagram of the structure of an electronic device according to one embodiment of the present disclosure is shown. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0020] In related technologies, the accuracy of volume prediction data based on machine learning algorithms is prone to significant fluctuations during peak e-commerce periods, resulting in a large deviation between the predicted data and the actual volume, making it difficult to provide effective guidance for site operations.

[0021] In related technologies, some unforeseen scenarios can completely alter existing logistics and transportation routes and arrangements. For example, natural disasters, traffic accidents, road closures for maintenance, or other emergencies may necessitate switching loading and unloading sites and changing existing logistics and transportation routes. In such cases, cargo volume forecasting methods based on fixed routes are insufficient to meet practical application needs.

[0022] In view of this, this application discloses a freight volume prediction method provided by an embodiment, which can accurately predict the allocated freight volume of each route and optimize the resource allocation of vehicles and sites.

[0023] It should be noted that, here and below, "route" primarily refers to routes related to logistics transportation, "vehicle" primarily refers to vehicles used for logistics transportation, and "site" primarily refers to sites related to logistics transportation (such as logistics transfer stations).

[0024] Please see Figure 1 The cargo volume forecasting method provided in one embodiment of this disclosure may include the following steps.

[0025] S1: Obtain historical waybill data for a predetermined time period, and generate a multi-dimensional initial probability transition matrix based on the historical waybill data. The initial probability transition matrix includes historical route resource information and route shipment probability information.

[0026] In this embodiment, historical waybill data may include historical waybill information and historical waybill routing information. Historical waybill information typically includes, but is not limited to, the following: waybill number, waybill generation time, shipping date and time, receiving date and time, shipper name, shipper identification information, loading address, consignee name, delivery address, freight amount, vehicle license plate number, cargo name, and cargo weight. Historical waybill routing information may involve the specific path and status of the goods during transportation, including but not limited to: waybill number, time of routing node occurrence, location of routing node occurrence, detailed description of routing node, routing node operation code, cargo status classification code, and cargo status classification name.

[0027] In this implementation, historical waybill data can be obtained through big data technology. Big data technology can employ techniques such as column storage, distributed computing, and parallel processing to improve data reading and processing efficiency. Big data technology can include ETL technology, which stands for Extract, Transform, and Load. This process extracts data from different sources, performs necessary transformations, and finally loads it into the target system (such as a data warehouse, data lake, or other analytics platform). Data transformation can include data purification techniques such as removing duplicate records, handling null values, detecting and handling outliers, and converting data types. Data transformation can also include data mapping and standardization techniques such as data format unification, date format standardization, and unit of measurement conversion. Furthermore, data transformation can include data quality checks such as verifying data integrity, consistency, and accuracy.

[0028] In this implementation, big data technology can be used to correlate and calculate massive amounts of historical waybills and their routes (hundreds of billions of data points) to generate an initial probability transition matrix with multiple dimensions, including schedules, routes, flow directions, products, and route delivery probabilities. The schedule and route information can constitute historical route resource information.

[0029] In this implementation, the initial probability transition matrix can be updated daily, and can be recalculated using historical waybill data within a predetermined time period from the update date. The predetermined time period can be based on days, such as a week, a month, or a quarter.

[0030] In some implementations, the step of acquiring historical waybill data for a predetermined time period and generating a multi-dimensional initial probability transition matrix includes: extracting target data from the historical waybill data that matches the target dimension, wherein the target dimension includes vehicle frequency, route code, origin, destination, product type or statistical shipment probability; and generating the initial probability transition matrix based on the target data.

[0031] Specifically, the initial probability transition matrix can contain information across multiple dimensions, such as location, shift, route code, cargo flow, product type, variance, coefficient of variation, and probability. Depending on the specific application scenario, the target dimensions may vary, but should at least include vehicle shifts, route codes, origin, destination, and statistical delivery probability as key dimensions for cargo volume prediction on each route. Explicit target dimensions can include all listed dimensions. In the initial probability transition matrix, the route delivery probability information can be normalized to ensure that the sum of the route delivery probabilities for all routes is 1, facilitating subsequent analysis and calculation.

[0032] In a practical application example, the initial probability transition matrix is ​​updated daily using a cover-over update method. The update assignment rule for the initial probability transition matrix is: daily, the initial probability transition matrix is ​​recalculated using waybill data from the past 30-day historical window, ensuring its reliability. The delivery process of a waybill can be divided into multiple route segments based on delivery nodes, such as 010R-010WB (origin-destination), connected by route codes. By statistically analyzing the segmented delivery patterns of historical multi-segment routes, the delivery probability for each shift and route segment can be generated.

[0033] It should be noted that, within the historical 30-day window of waybill data used daily, for signed-for waybills, the completion time period and all routing information can be retrieved using the waybill number. For unsigned waybills, the routing inference engine can be used to infer the routing information for the incomplete parts to complete the routing information. After this step, all waybills from the past 30 days can be considered as signed-for waybills.

[0034] In a practical application example, the initial probability transition matrix can characterize the statistical regularity of goods being sent via each route. An initial probability transition matrix before updating can be as follows:

[0035]

[0036] In this practical application example, the updated initial probability transition matrix can be as follows:

[0037]

[0038]

[0039] As can be seen, the updated initial probability transition matrix contains changes in the route delivery probability information compared to the original initial probability transition matrix, thus ensuring that the initial probability transition matrix is ​​more effective and accurate, and making the final predicted expected allocation of cargo volume more accurate.

[0040] In some implementations, the initial probability transition matrix can be transmitted to a big data online analysis and processing database for further processing.

[0041] Specifically, an Online Analytical Processing (OLAP) database is a relational database system specifically designed for processing and analyzing large-scale datasets. Based on a multidimensional data model, it supports online analytical processing, providing users with fast, coherent, and multi-dimensional data analysis capabilities. Using an OLAP database, the initial probability transition matrix can be effectively stored, updated, and adjusted.

[0042] S2: Obtain real-time waybill barcode operation data, and based on the real-time waybill barcode operation data, determine the planned route resource information, real-time vehicle status, and real-time cargo volume summary data for the day.

[0043] In this embodiment, a big data online analysis and processing database can be used to effectively store and analyze real-time waybill barcode operation data, thereby extracting the daily planned route resource information, real-time vehicle status, and real-time freight volume summary data contained in the real-time waybill barcode operation data. The daily planned route resource information may include the day's schedule information and route information.

[0044] In some implementations, a message queue framework can be used to obtain real-time waybill / barrel operation data. After data cleaning, the cleaned real-time waybill / barrel operation data can be stored in an analysis database. In the analysis database, based on the operation codes contained in the real-time waybill / barrel operation data, the planned route resource information, real-time vehicle status, and real-time freight volume summary data for the day can be determined.

[0045] Specifically, the database being analyzed can be a big data online analytics database, such as the DORIS database. Barcode scanning can generate different operation codes at different stages of the logistics process. For example, loading, unloading, dispatching, arrival, receiving, delivery, and warehousing can all correspond to different operation codes, which can be in the form of numbers, letters, characters, or one or more other formats.

[0046] In a practical application example, a message queue framework, data cleaning tools, and a big data online analysis and processing database can be considered a classic three-piece set for handling big data distributed real-time computing tasks, corresponding to data extraction, data transformation, and data loading tasks, respectively.

[0047] Message queue frameworks are primarily used to store large amounts of real-time messages, such as barcode scanner data for all waybills across the network. Each waybill generates scanning data at every stage of its operation. This network-wide barcode scanner data amounts to billions of data points daily, which can be stored in real-time using a message queue framework.

[0048] Data cleaning tools are mainly used to clean, correlate, and calculate real-time data to generate relevant statistical indicators.

[0049] The big data online analysis and processing database is mainly used to store cleaned and calculated data, and can be used for storing and querying the results.

[0050] S3: Based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, adjust the route delivery probability information of the initial probability transition matrix to obtain the target probability transition matrix.

[0051] In this implementation, the planned route resources for the day are combined with real-time vehicle status and compared with historical route resources. Data processing operations such as filtering, probability recalculation, and normalization of waybill data can be performed at the route level to adjust the route delivery probability information of the initial probability transition matrix.

[0052] In some implementations, adjusting the route delivery probability information of the initial probability transition matrix based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status includes: determining a first route based on the planned route resource information for the day and the historical route resource information, adding matrix information corresponding to the first route to the initial probability transition matrix, and setting the route delivery probability corresponding to the first route to zero; determining a second route based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, and setting the route delivery probability corresponding to the second route to zero in the initial probability transition matrix; determining a third route based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, and normalizing the route delivery probability corresponding to the third route in the initial probability transition matrix; wherein, the first route represents a newly added route for the day, the second route represents a historical route with no vehicles currently waiting to depart, and the third route represents a historical route with vehicles currently waiting to depart.

[0053] Specifically, the available route resources for the day are compared with historical route resources. If a route has never existed before, it is considered a new route. If there is no historical probability data for a new route, its shipping probability can be set to 0, and it will not be included in subsequent cargo volume forecasts.

[0054] Compare the current day's route resources with historical route resources. If a route exists historically, then for that route, it's necessary to obtain in real-time whether there are any undeparted vehicles waiting to depart. If there are no vehicles waiting to depart, the current number of vehicles waiting to depart is 0; if there are vehicles waiting to depart, the current number of vehicles waiting to depart is greater than 0. If the number of vehicles waiting to depart is 0, it means there are no available vehicles on that route, and the original probability for that route is set to 0, and it will no longer be included in subsequent freight volume forecasts.

[0055] After addressing the two special cases mentioned above, the remaining route resources need to have their probabilities recalculated and split. By standardizing the unit of the delivery probability for each route, the sum of the delivery probabilities for each route is guaranteed to equal 1. This ensures that the cargo volume can be fully split during cargo volume forecasting, avoiding a decrease in the accuracy of cargo volume forecasting due to non-standardization issues caused by inconsistent data units.

[0056] During the adjustment of the initial probability transition matrix, the first route can be determined first, and its corresponding matrix information (such as vehicle schedule, route code, origin, destination, and statistical delivery probability) can be added to the initial probability transition matrix. Since the first route is a newly added route, its delivery probability is initially set to zero. Subsequently, the matrix information for the second route can be updated. Because the second route represents a historical route with no vehicles currently waiting to depart, its delivery probability also needs to be set to zero. Finally, for the remaining third route, its delivery probability is normalized to ensure that the sum of the delivery probabilities of all routes in the target probability transition matrix remains 1.

[0057] S4: Based on the real-time cargo volume summary data and the target probability transition matrix, determine the expected cargo volume allocation for each target route.

[0058] In this embodiment, by utilizing the target delivery probability of each route in the target probability transition matrix, the total delivery volume of each shift can be allocated to each route, and the expected allocation volume of each target route can be predicted.

[0059] In some implementations, determining the expected allocation of cargo volume for each target route based on the real-time cargo volume summary data and the target probability transition matrix includes: determining the shift cargo volume data for each shift based on the real-time cargo volume summary data; determining the target delivery probability for each target route based on the target probability transition matrix; and determining the expected allocation of cargo volume for the target route based on the shift cargo volume data and the target delivery probability.

[0060] In a practical application example, the expected cargo volume allocated to each target route is the product of the shift cargo volume of each target route and the route delivery probability of each target route. That is, the method for determining the expected cargo volume allocated to each target route is: Shift cargo volume * Route delivery probability = Expected cargo volume allocated to the route. Shift cargo volume represents the total cargo volume to be allocated for each shift, and the route delivery probability represents the target delivery probability of different routes under each shift in the target probability transition matrix.

[0061] In some implementations, after determining the expected allocation volume for each target route, the method further includes: comparing the expected allocation volume with the total carrying capacity of the candidate vehicles; if the ratio of the expected allocation volume to the total carrying capacity is greater than a full load threshold, then outputting a full load warning message; if the ratio of the expected allocation volume to the total carrying capacity is less than an underload threshold, then outputting an underload warning message.

[0062] Specifically, the full load threshold needs to be greater than the low load threshold. For example, the full load threshold can exceed a preset value of the low load threshold. Or, for example, the full load threshold can be a preset multiple of the low load threshold.

[0063] Based on the expected cargo volume, for routes with the risk of overloading, a request to add vehicles can be initiated in advance to increase vehicle resources on routes with the risk of overloading in order to accurately deal with the risk of full load; for routes with insufficient cargo volume, a request to reduce vehicles can be initiated in advance to reduce vehicle resources on routes with insufficient cargo volume, thereby enabling vehicle resources to be transferred to where they are needed and avoiding vehicle resources being idle and thus avoiding waste of vehicle resources.

[0064] In some implementations, after determining the expected allocation volume for each target route, the method further includes: determining a preliminary transportation plan based on the expected allocation volume; determining the preliminary transportation plan as a slow train transportation plan when the expected allocation volume is lower than a first threshold; determining the preliminary transportation plan as an express train transportation plan when the expected allocation volume is higher than the first threshold and lower than a second threshold; determining the preliminary transportation plan as a direct delivery transportation plan when the expected allocation volume is higher than the second threshold; and connecting the preliminary transportation plan with an operations research algorithm to determine the target transportation plan.

[0065] Specifically, in a route with multiple transfer stations, a slow train transportation plan might stop at every transfer station; an express train transportation plan might stop at only some transfer stations; and a direct train transportation plan might run directly from the origin to the destination without stopping at any transfer stations. Based on the expected cargo volume, a better initial transportation plan can be recommended to improve overall resource utilization. Subsequently, using operations research algorithms, the initial transportation plan can be further optimized to determine the target transportation plan and ensure its feasibility.

[0066] Please see Figure 2 This application discloses a cargo volume forecasting method according to one embodiment, which may include the following steps.

[0067] Step S21: Using big data technology, correlate and calculate data from hundreds of billions of historical waybills and waybill routes to generate an initial probability transition matrix for the dimensions of shifts, routes, directions, and products. The data update strategy for the initial probability transition matrix is ​​to update it in detail once a day. The matrix information of the initial probability transition matrix can cover information such as location, shift, route code, direction, product type, variance, coefficient of variation, and probability.

[0068] Step S22: Extract the generated initial probability transition matrix data into the big data online analysis and processing database using ETL technology.

[0069] Step S23: Using a data cleaning tool, the real-time waybill barcode operation data is read from the message queue framework, and specific operation codes are selected for aggregation to generate real-time cargo volume summary data in the OD (Origin-Destination) dimension.

[0070] Step S24: Combine the planned route resources for the day with the real-time vehicle status, compare with historical route resources, filter the route data, recalculate the probability, and normalize the data.

[0071] Step S25: Call the allocation algorithm every 10 minutes to split the real-time cargo volume into different lines according to probability. The cargo volume to be shipped can be predicted 6 hours, 5 hours... 1 hour and 0 hours in advance.

[0072] Step S26: Based on the full load status of the remaining vehicles, output low load and full load alarms for the vehicles in real time.

[0073] The projected cargo volume allocated to each target route in this plan can be the expected cargo volume allocated to each target route within a preset future time period (e.g., 6 hours, 5 hours... 1 hour, 0 hours, etc.). Based on changes in the projected cargo volume, human resources can be dynamically allocated to minimize cargo dwell time on the site, improve transshipment efficiency, and reduce the risk of cargo being left unattended. In emergency situations, real-time, dynamically changing projected cargo volume data can provide crucial reference, allowing for advance planning to salvage tickets on-time. Based on changes in future hourly cargo volume, empty warehouse resources can be identified, and remaining warehouse space can be sold to improve vehicle resource utilization.

[0074] In some implementations, the predicted cargo volume can be stored every 10 minutes and compared with the actual cargo volume to generate a T-1 report, which is used to evaluate the accuracy of the projection and alarm, and facilitates guidance to improve the accuracy of the predicted cargo volume.

[0075] This application discloses a technical solution provided by one or more embodiments, which can acquire a large amount of historical waybill data, establish a multi-dimensional initial probability transition matrix, and ensure the reliability of the prediction method. Through real-time waybill barcode scanner operation data, the route resource information, vehicle status information, and freight volume summary information for the day can be determined. Based on this, and using the planned route resource information for the day, historical route resource information, and real-time vehicle status, the probability distribution of the initial probability transition matrix can be optimized and adjusted to obtain the target probability transition matrix, ensuring the real-time performance and accuracy of the prediction method. Using real-time freight volume summary data and the target probability transition matrix, the expected freight volume allocated to each target route can be accurately predicted, facilitating the determination of the expected arrival freight volume at each site, thereby optimizing the resource allocation of vehicles and sites.

[0076] This application discloses a technical solution provided by one or more embodiments, which can accurately predict the expected freight volume allocated to each target route, thereby improving the utilization efficiency of route resources. Based on the predicted freight volume data, vehicle resources can be dynamically and more rationally allocated, making adding and removing vehicles more precise and efficient. The expected freight volume allocated to each target route also helps determine the expected arrival volume of freight at each site, which can reduce the operating costs of the sites.

[0077] This application discloses a technical solution provided by one or more embodiments, which can output full load and low load alarms in advance based on predicted cargo volume data, and can give the site sufficient time to handle abnormal situations, thereby reducing the risk of cargo falling off the site and improving customer satisfaction.

[0078] This application discloses a technical solution provided by one or more embodiments, which can be connected to operations planning algorithms based on real-time predicted cargo volume data at the route dimension, and can perform fast and slow separation loading in advance according to the predicted cargo volume, thereby reducing the difficulty of downstream unloading operations.

[0079] It should be noted that, in some practical application tests, the cargo volume prediction method provided by one or more embodiments of this disclosure can improve the accuracy of cargo volume prediction by an average of 15% compared with the accuracy of prediction by machine learning models.

[0080] Please see Figure 3 This disclosure also provides a cargo volume prediction device, the device comprising:

[0081] The matrix building unit 100 is used to acquire historical waybill data for a predetermined time period and generate a multi-dimensional initial probability transition matrix based on the historical waybill data. The initial probability transition matrix includes historical route resource information and route shipment probability information.

[0082] The data acquisition unit 200 is used to acquire real-time waybill barcode operation data, and based on the real-time waybill barcode operation data, determine the planned route resource information, real-time vehicle status and real-time cargo volume summary data for the day;

[0083] The matrix update unit 300 is used to adjust the route delivery probability information of the initial probability transition matrix based on the planned route resource information of the day, the historical route resource information and the real-time vehicle status, so as to obtain the target probability transition matrix.

[0084] The cargo volume prediction unit 400 is used to determine the expected cargo volume allocation for each target route based on the real-time cargo volume summary data and the target probability transition matrix.

[0085] In one implementation, the matrix building unit 100 is specifically used to extract target data from the historical waybill data that matches the target dimension, the target dimension including vehicle frequency, route code, origin, destination, product type or statistical shipment probability; and generate the initial probability transition matrix based on the target data.

[0086] In one embodiment, the data acquisition unit 200 is specifically used to: acquire the real-time waybill barcode operation data based on a message queue framework; perform data cleaning on the real-time waybill barcode operation data and store the cleaned real-time waybill barcode operation data into an analysis database; and in the analysis database, determine the planned route resource information for the day, the real-time vehicle status, and the real-time freight volume summary data based on the operation codes contained in the real-time waybill barcode operation data.

[0087] In one implementation, the matrix update unit 300 is specifically configured to: determine a first route based on the planned route resource information for the day and the historical route resource information, add matrix information corresponding to the first route to the initial probability transition matrix, and set the delivery probability of the route corresponding to the first route to zero; determine a second route based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, and set the delivery probability of the route corresponding to the second route to zero in the initial probability transition matrix; determine a third route based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, and normalize the delivery probability of the route corresponding to the third route in the initial probability transition matrix; wherein, the first route represents a newly added route for the day, the second route represents a historical route with no vehicles currently waiting to depart, and the third route represents a historical route with vehicles currently waiting to depart.

[0088] In one embodiment, the cargo volume prediction unit 400 is specifically used to: determine the cargo volume data of each shift based on the real-time cargo volume summary data; determine the target delivery probability of each target route based on the target probability transition matrix; and determine the expected allocated cargo volume of the target route based on the shift cargo volume data and the target delivery probability.

[0089] In one embodiment, the cargo volume prediction device further includes an early warning unit 500. The early warning unit 500 is used to compare the expected allocated cargo volume with the total carrying capacity of the candidate vehicles; if the ratio of the expected allocated cargo volume to the total carrying capacity is greater than a full-load threshold, a full-load warning message is output; if the ratio of the expected allocated cargo volume to the total carrying capacity is less than an underload threshold, an underload warning message is output.

[0090] In one embodiment, the cargo volume prediction device further includes a planning unit 600. The planning unit 600 is configured to: determine a preliminary transportation plan based on the expected cargo volume; determine the preliminary transportation plan as a slow-speed transportation plan when the expected cargo volume is lower than a first threshold; determine the preliminary transportation plan as an express-speed transportation plan when the expected cargo volume is higher than the first threshold and lower than a second threshold; determine the preliminary transportation plan as a direct-delivery transportation plan when the expected cargo volume is higher than the second threshold; and connect the preliminary transportation plan to an operations research algorithm to determine a target transportation plan.

[0091] The various units described in the above embodiments can be implemented by a computer chip or by a product with a certain function. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0092] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0093] Please see Figure 4 This disclosure also provides an electronic device, which includes a memory and a processor. The memory is used to store a computer program, which, when executed by the processor, implements the above-described cargo volume prediction method.

[0094] This disclosure also provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the above-described cargo volume prediction method.

[0095] The processor can be a central processing unit (CPU). It can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof.

[0096] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby implementing the methods in the above-described embodiments.

[0097] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0098] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0099] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, embodiments of apparatus, devices, and storage media are basically similar to method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0100] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

[0101] Although embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A cargo volume prediction method characterized by, The method includes: Obtain historical waybill data for a predetermined time period, and generate a multi-dimensional initial probability transition matrix based on the historical waybill data. The initial probability transition matrix includes historical route resource information and route shipment probability information. Acquire real-time waybill barcode operation data, and based on the real-time waybill barcode operation data, determine the planned route resource information, real-time vehicle status, and real-time freight volume summary data for the day; Based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, the route delivery probability information of the initial probability transition matrix is ​​adjusted to obtain the target probability transition matrix. Based on the real-time cargo volume summary data and the target probability transition matrix, the expected cargo volume allocation for each target route is determined.

2. The method of claim 1, wherein, The generation of a multi-dimensional initial probability transition matrix based on the historical waybill data includes: Extract target data that matches the target dimension from the historical waybill data. The target dimension includes vehicle frequency, route code, origin, destination, product type, or statistical shipment probability. Based on the target data, the initial probability transition matrix is ​​generated.

3. The method of claim 1, wherein, The process of acquiring real-time waybill barcode scanner operation data and determining the planned route resource information, real-time vehicle status, and real-time freight volume summary data for the day based on the real-time waybill barcode scanner operation data includes: The real-time waybill barcode operation data is obtained based on a message queue framework; The real-time waybill barcode operation data is cleaned, and the cleaned real-time waybill barcode operation data is stored in the analysis database. In the analysis database, the planned route resource information for the day, the real-time vehicle status, and the real-time freight volume summary data are determined based on the operation codes contained in the real-time waybill barcode operation data.

4. The method of claim 1, wherein, The step of adjusting the route delivery probability information of the initial probability transition matrix based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status includes: Based on the planned route resource information for the day and the historical route resource information, a first route is determined, and the matrix information corresponding to the first route is added to the initial probability transition matrix, and the shipping probability of the route corresponding to the first route is set to zero. Based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, a second route is determined, and the delivery probability of the route corresponding to the second route is set to zero in the initial probability transition matrix. Based on the planned route resource information for the day, the historical route resource information, and the real-time vehicle status, a third route is determined, and the delivery probability of the route corresponding to the third route is normalized in the initial probability transition matrix. The first route represents a newly added route on the current day, the second route represents a historical route with no vehicles currently waiting to depart, and the third route represents a historical route with vehicles currently waiting to depart.

5. The method according to claim 1, characterized in that, The step of determining the expected allocation of cargo volume for each target route based on the real-time cargo volume summary data and the target probability transition matrix includes: Based on the real-time cargo volume summary data, determine the cargo volume data for each shift; Based on the target probability transition matrix, the target shipment probability of each target route is determined; Based on the freight volume data of the scheduled trips and the target delivery probability, the expected freight volume allocated to the target route is determined.

6. The method according to any one of claims 1 to 5, characterized in that, After determining the expected allocation of cargo volume for each target route, the method further includes: Compare the expected cargo volume with the total carrying capacity of the candidate vehicles; If the ratio of the expected allocated cargo volume to the total carrying capacity is greater than the full load threshold, a full load warning message will be output. If the ratio of the expected allocated cargo volume to the total carrying capacity is less than the low load threshold, a low load warning message will be output.

7. The method according to any one of claims 1 to 5, characterized in that, After determining the expected allocation of cargo volume for each target route, the method further includes: Based on the expected allocation of cargo volume, a preliminary transportation plan is determined; When the expected allocation volume is lower than the first threshold, the initial selected transportation plan is determined to be the slow vehicle transportation plan; When the expected allocation volume is higher than the first threshold and lower than the second threshold, the initial selected transportation plan is determined to be the express transportation plan. When the expected allocation volume is higher than the second threshold, the initial selected transportation plan is determined to be a direct delivery transportation plan; The initially selected transportation plan is then used in conjunction with an operations research algorithm to determine the target transportation plan.

8. A cargo volume prediction device, characterized in that, The device includes: The matrix building unit is used to acquire historical waybill data for a predetermined time period and generate a multi-dimensional initial probability transition matrix based on the historical waybill data. The initial probability transition matrix includes historical route resource information and route shipment probability information. The data acquisition unit is used to acquire real-time waybill barcode operation data, and based on the real-time waybill barcode operation data, determine the planned route resource information, real-time vehicle status and real-time freight volume summary data for the day; The matrix update unit is used to adjust the route delivery probability information of the initial probability transition matrix based on the planned route resource information of the day, the historical route resource information and the real-time vehicle status, so as to obtain the target probability transition matrix. The cargo volume prediction unit is used to determine the expected cargo volume allocation for each target route based on the real-time cargo volume summary data and the target probability transition matrix.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory being used to store a computer program that, when executed by the processor, implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images