Method and device for generating vehicle scheduling scheme and electronic equipment

By extracting historical regular cargo volume sequences from the original cargo volume sequence, training a cargo volume prediction model, and generating a vehicle scheduling scheme, the problem of low loading rate of self-operated vehicles was solved, achieving higher loading rate and lower transportation cost.

CN118941053APending Publication Date: 2024-11-12BEIJING JINGDONG QIANSHITECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411161752.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The low loading rate of self-operated vehicles in existing technologies leads to wasted transport capacity and high transportation costs, mainly due to inaccurate transportation demand forecasting and the inability to flexibly respond to dynamic changes in routes and special circumstances.

Method used

By extracting historical regular freight volume sequences from the original freight volume sequence, a freight volume prediction model is trained to predict the freight volume sequence of future transportation routes. Combined with a greedy algorithm, a vehicle scheduling scheme is generated to optimize vehicle utilization and reduce transportation costs.

Benefits of technology

It improved the load factor of transportation, reduced transportation costs, and enabled more accurate transportation demand forecasting and flexible vehicle scheduling.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941053A_ABST
    Figure CN118941053A_ABST
Patent Text Reader

Abstract

The invention discloses a method and device for generating a vehicle scheduling scheme and electronic equipment, and relates to the technical field of logistics distribution. The method comprises the following steps: acquiring an original goods quantity sequence of a transportation line in a past period of time, and extracting a historical conventional goods quantity sequence from the original goods quantity sequence; training by adopting the historical conventional goods quantity sequence to obtain a goods quantity prediction model corresponding to the transportation line, and predicting a predicted conventional goods quantity sequence of the transportation line in a period of time in the future by adopting the goods quantity prediction model; according to the predicted conventional cargo quantity sequence and the corresponding fluctuation cargo quantity sequence, generating a predicted cargo quantity sequence of the transportation line in a period of time in the future; and generating a vehicle scheduling scheme according to the predicted cargo quantity sequence of each line. According to the embodiment, the technical problems of transport capacity waste and high transport cost caused by low vehicle loading rate can be solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of logistics distribution, and in particular to a method, device and electronic equipment for generating a vehicle dispatching plan. Background Art

[0002] Currently, there are two types of vehicles that perform transportation tasks: third-party vehicles and self-operated vehicles. Third-party vehicles are also called outsourced vehicles or leased vehicles. They are vehicles owned and operated by independent transportation companies, logistics service providers or individuals. They reach an agreement with the shipper and consignee of the goods to provide transportation services, while self-operated vehicles are vehicles owned by the company.

[0003] Generally speaking, the cost of a self-owned vehicle to perform a single transport task is lower than that of a third-party vehicle. Purchasing self-owned vehicles to replace third-party vehicles can reduce the company's transportation costs. However, the problem faced by self-owned vehicles is the lack of flexibility. If there is no properly arranged transport task, the self-owned vehicle may be idle for a long time. Its depreciation and employee wages may lead to poor cost reduction effects when replacing third-party vehicles with self-owned vehicles.

[0004] At present, the business system's method of predicting future transportation demand is mainly based on rule-based algorithms. These algorithms use rules to predict the future transportation demand of a certain route, such as judging whether the frequency of past transportation tasks exceeds a certain threshold. If so, it is considered that there is a stable transportation demand for this route every day. After analyzing the stable transportation demand of all routes, the current unified dispatching algorithm will traverse and mine all loops through depth-first search, find the loop route with the highest third-party vehicle transportation cost in historical records, and use self-operated vehicles to perform the transportation tasks of this route.

[0005] Although existing rule-based algorithms are simple and easy to use in some cases, they have shortcomings, mainly manifested in that self-operated vehicles often have abnormally low loading rates, which not only causes a waste of transportation capacity but also directly leads to an increase in transportation costs. Summary of the invention

[0006] In view of this, an embodiment of the present invention provides a method, device and electronic device for generating a vehicle dispatching plan to solve the technical problem that low vehicle loading rate leads to waste of transportation capacity and high transportation cost.

[0007] To achieve the above object, according to one aspect of an embodiment of the present invention, a method for generating a vehicle dispatching plan is provided, comprising:

[0008] Obtaining an original cargo volume sequence of a transport route in the past period of time, and extracting a historical regular cargo volume sequence from the original cargo volume sequence;

[0009] The historical conventional cargo volume sequence is used to train a cargo volume prediction model corresponding to the transport route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in the future;

[0010] Generate a predicted cargo volume sequence for the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence;

[0011] Generate a vehicle dispatch plan based on the predicted cargo volume sequence of each route.

[0012] Optionally, extracting a historical conventional cargo volume sequence from the original cargo volume sequence includes:

[0013] Extracting a characteristic cargo quantity sequence from the original cargo quantity sequence, thereby obtaining a non-characteristic cargo quantity sequence after removing the characteristic cargo quantity sequence;

[0014] Generate a characteristic effect cargo volume sequence and a periodic cargo volume sequence according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence;

[0015] The periodic cargo volume sequence is subtracted from the non-characteristic cargo volume sequence to obtain the historical regular cargo volume sequence.

[0016] Optionally, generating a characteristic effect sequence and a periodic cargo volume sequence according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence includes:

[0017] Performing linear interpolation processing on the non-characteristic cargo quantity sequence to obtain an interpolated cargo quantity sequence;

[0018] Subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence to obtain a characteristic effect sequence;

[0019] Decomposing the non-characteristic cargo volume sequence according to different period dimensions and using a time series decomposition algorithm, thereby obtaining a sub-period cargo volume sequence corresponding to each period dimension;

[0020] Generate a periodic cargo quantity sequence according to the sub-period cargo quantity sequences corresponding to each periodic dimension.

[0021] Optionally, generating a periodic cargo quantity sequence according to the sub-period cargo quantity sequences corresponding to the respective periodic dimensions includes:

[0022] Add the sub-period cargo volume sequences corresponding to each period dimension to obtain a period cargo volume sequence;

[0023] The various periodic dimensions include the year dimension, the day dimension under the month dimension, and the day dimension under the week dimension.

[0024] Optionally, the cargo volume prediction model corresponding to the transport route is obtained by training the historical conventional cargo volume sequence, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in a future period of time, including:

[0025] Determine each historical duration according to a preset duration interval and a preset step length;

[0026] For each historical duration, a cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence corresponding to the historical duration, and the average absolute percentage error of the cargo volume prediction model is calculated;

[0027] Filter out the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration;

[0028] The cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration, as well as the preset probability quantile, is used to forecast the predicted regular cargo volume sequence of the transport route in the future.

[0029] Optionally, a cargo volume prediction model corresponding to the transport route is obtained by training a historical conventional cargo volume sequence corresponding to the historical duration, and a mean absolute percentage error of the cargo volume prediction model is calculated, including:

[0030] The log-normal distribution model is fitted using the historical conventional cargo volume sequence corresponding to the historical duration, so as to train and obtain a cargo volume prediction model corresponding to the transportation route;

[0031] Using the cargo volume forecasting model to forecast a forecast cargo volume sequence for a future period of time corresponding to the historical time length;

[0032] The average absolute percentage error of the cargo volume prediction model is calculated based on the predicted cargo volume sequence in the future period corresponding to the historical time length and the actual cargo volume sequence in the future period corresponding to the historical time length.

[0033] Optionally, calculating the mean absolute percentage error of the cargo volume prediction model according to the predicted cargo volume sequence in the future period corresponding to the historical duration and the actual cargo volume sequence in the future period corresponding to the historical duration includes:

[0034] For each time point in the future period corresponding to the historical time length, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point; if so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error; if not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume;

[0035] The average absolute percentage error of the cargo volume prediction model is calculated according to the absolute percentage errors corresponding to the respective time points and / or the absolute percentage errors after reducing the penalty.

[0036] Optionally, generating a predicted cargo volume sequence of the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence includes:

[0037] Add the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain a fluctuating cargo volume sequence;

[0038] Extracting a fluctuating cargo volume sequence corresponding to the predicted regular cargo volume sequence from the fluctuating cargo volume sequence;

[0039] The predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence are added together to obtain the predicted cargo volume sequence of the transport route in a future period of time.

[0040] Optionally, a vehicle dispatching plan is generated based on the predicted cargo volume sequence of each route, including:

[0041] According to the predicted cargo volume sequence of each route, each route is matched with a plurality of first vehicles and a plurality of second vehicles one by one;

[0042] According to the transportation cost of the first vehicle and the transportation cost of the second vehicle, a vehicle dispatching scheme with the greatest reduction in total transportation cost is solved by using a greedy algorithm.

[0043] In addition, according to another aspect of an embodiment of the present invention, there is provided a device for generating a vehicle dispatching plan, comprising:

[0044] The first sequence module is used to obtain the original cargo volume sequence of the transportation route in the past period of time, and extract the historical regular cargo volume sequence from the original cargo volume sequence;

[0045] A prediction module, used to train the historical conventional cargo volume sequence to obtain a cargo volume prediction model corresponding to the transport route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in the future;

[0046] A second sequence module is used to generate a predicted cargo volume sequence for the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence;

[0047] The generation module is used to generate a vehicle dispatching plan based on the predicted cargo volume sequence of each route.

[0048] Optionally, the first sequence module is further used for:

[0049] Extracting a characteristic cargo quantity sequence from the original cargo quantity sequence, thereby obtaining a non-characteristic cargo quantity sequence after removing the characteristic cargo quantity sequence;

[0050] Generate a characteristic effect cargo volume sequence and a periodic cargo volume sequence according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence;

[0051] The periodic cargo volume sequence is subtracted from the non-characteristic cargo volume sequence to obtain the historical regular cargo volume sequence.

[0052] Optionally, the first sequence module is further used for:

[0053] Performing linear interpolation processing on the non-characteristic cargo quantity sequence to obtain an interpolated cargo quantity sequence;

[0054] Subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence to obtain a characteristic effect sequence;

[0055] Decomposing the non-characteristic cargo volume sequence according to different period dimensions and using a time series decomposition algorithm, thereby obtaining a sub-period cargo volume sequence corresponding to each period dimension;

[0056] Generate a periodic cargo quantity sequence according to the sub-period cargo quantity sequences corresponding to each periodic dimension.

[0057] Optionally, the first sequence module is further used for:

[0058] Add the sub-period cargo volume sequences corresponding to each period dimension to obtain a period cargo volume sequence;

[0059] The various periodic dimensions include the year dimension, the day dimension under the month dimension, and the day dimension under the week dimension.

[0060] Optionally, the prediction module is further used for:

[0061] Determine each historical duration according to a preset duration interval and a preset step length;

[0062] For each historical duration, a cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence corresponding to the historical duration, and the average absolute percentage error of the cargo volume prediction model is calculated;

[0063] Filter out the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration;

[0064] The cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration, as well as the preset probability quantile, is used to forecast the predicted regular cargo volume sequence of the transport route in the future.

[0065] Optionally, the prediction module is further used for:

[0066] The log-normal distribution model is fitted using the historical conventional cargo volume sequence corresponding to the historical duration, so as to train and obtain a cargo volume prediction model corresponding to the transportation route;

[0067] Using the cargo volume forecasting model to forecast a forecast cargo volume sequence for a future period of time corresponding to the historical time length;

[0068] The average absolute percentage error of the cargo volume prediction model is calculated based on the predicted cargo volume sequence in the future period corresponding to the historical time length and the actual cargo volume sequence in the future period corresponding to the historical time length.

[0069] Optionally, the prediction module is further used for:

[0070] For each time point in the future period corresponding to the historical time length, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point; if so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error; if not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume;

[0071] The average absolute percentage error of the cargo volume prediction model is calculated according to the absolute percentage errors corresponding to the respective time points and / or the absolute percentage errors after reducing the penalty.

[0072] The second sequence module is further used for:

[0073] Add the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain a fluctuating cargo volume sequence;

[0074] Extracting a fluctuating cargo volume sequence corresponding to the predicted regular cargo volume sequence from the fluctuating cargo volume sequence;

[0075] The predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence are added together to obtain the predicted cargo volume sequence of the transport route in a future period of time.

[0076] Optionally, the generating module is further used for:

[0077] According to the predicted cargo volume sequence of each route, each route is matched with a plurality of first vehicles and a plurality of second vehicles one by one;

[0078] According to the transportation cost of the first vehicle and the transportation cost of the second vehicle, a vehicle dispatching scheme with the greatest reduction in total transportation cost is solved by using a greedy algorithm.

[0079] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, including:

[0080] one or more processors;

[0081] a storage device for storing one or more programs,

[0082] When the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any one of the above embodiments.

[0083] According to another aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, the method described in any one of the above embodiments is implemented.

[0084] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method described in any one of the above embodiments is implemented.

[0085] An embodiment of the above invention has the following advantages or beneficial effects: because the historical conventional cargo volume sequence is extracted from the original cargo volume sequence, and the cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence, so that the cargo volume prediction model is used to predict the predicted conventional cargo volume sequence of the transportation route in the future period of time, and the predicted cargo volume sequence of the transportation route in the future period of time is generated according to the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence, so the technical problem of low vehicle loading rate leading to waste of transportation capacity and high transportation cost in the prior art is overcome. The embodiment of the present invention can accurately predict the cargo volume of each transportation route in the future period of time, and generate a vehicle scheduling plan with the lowest transportation cost based on the predicted cargo volume, which not only improves the loading rate but also reduces the transportation cost.

[0086] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0088] Figure 1 is a flow chart of a method for generating a vehicle dispatching plan according to an embodiment of the present invention;

[0089] Figure 2is a flow chart of a method for generating a vehicle dispatching plan according to a reference embodiment of the present invention;

[0090] Figure 3 is a comparison chart of cargo volume prediction results according to an embodiment of the present invention;

[0091] Figure 4 is a flow chart of a method for generating a vehicle dispatching plan according to another reference embodiment of the present invention;

[0092] Figure 5 is a schematic diagram of an apparatus for generating a vehicle dispatching plan according to an embodiment of the present invention;

[0093] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0094] Figure 7 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0095] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0096] It should be noted that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of user personal information involved in the technical solution of this disclosure are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information, network security and national security.

[0097] The reason for the low loading rate of transport coordination and scheduling is that the first is inaccurate forecast of transport demand. Predicting future transport task demand based on the dispatch frequency of historical transport tasks is based on simple rules. This rule algorithm based on static thresholds is usually unable to capture the complex dynamic changes and potential nonlinear relationships of route cargo volume. At the same time, it often lacks sufficient sensitivity to some predictable information in the future. For example, for routes dealing with special circumstances (such as holidays, special events, etc.), the rule algorithm cannot accurately handle this part of the request, resulting in an inability to accurately predict cargo volume.

[0098] In the case of inaccurate predictions, the current transportation scheduling based on transportation tasks lacks flexibility. The transportation tasks predicted for each route are fixed (for example, the existing method will assume that a certain route will have a demand for the shipment of a 9.6-meter truck every day for the next month, and this will not change). This makes it impossible for the algorithm to adapt to the specific needs and changes of different routes when making automatic decisions, nor can it automatically adjust thresholds or adapt to new transportation modes. Therefore, when faced with new transportation needs or changes in the transportation environment, it is impossible to generate a suitable vehicle scheduling plan, resulting in waste of transportation capacity and high transportation costs.

[0099] Moreover, the parameters of the rule-based algorithm mainly rely on manually defined rules and empirical judgments. The current rule-based algorithm cannot make full use of historical data. After manual decision-making, no uncertainty is introduced, which makes it easy for low loading to occur, and the generated vehicle scheduling plan is not optimized enough.

[0100] Figure 1 is a flow chart of a method for generating a vehicle dispatching scheme according to an embodiment of the present invention. As an embodiment of the present invention, Figure 1 As shown, the method for generating a vehicle dispatching plan may include:

[0101] Step 101, obtaining the original cargo volume sequence of the transport route in the past period of time, and extracting the historical regular cargo volume sequence from the original cargo volume sequence.

[0102] First, obtain the daily waybill data of each transport route in the past period of time (such as the past year, the past two years, the past three years or the past five years, etc.), and then aggregate them by day to obtain the daily historical cargo volume of each transport route, and then fill the missing data of the aggregated historical cargo volume to obtain the original cargo volume sequence of each transport route, and finally extract the historical conventional cargo volume sequence of each transport route from the original cargo volume sequence of each transport route. The embodiment of the present invention uses the historical conventional cargo volume sequence to train the cargo volume prediction model, which can improve the prediction accuracy of the model.

[0103] Optionally, extracting the historical conventional cargo quantity sequence from the original cargo quantity sequence includes: extracting the characteristic cargo quantity sequence from the original cargo quantity sequence, thereby obtaining a non-characteristic cargo quantity sequence after excluding the characteristic cargo quantity sequence; generating a characteristic effect cargo quantity sequence and a periodic cargo quantity sequence according to the characteristic cargo quantity sequence and the non-characteristic cargo quantity sequence; and subtracting the periodic cargo quantity sequence from the non-characteristic cargo quantity sequence to obtain the historical conventional cargo quantity sequence. In order to accurately extract the historical conventional cargo quantity sequence from the original cargo quantity sequence, first, according to each characteristic date in the past period of time, extract the characteristic cargo quantity sequence from the original cargo quantity sequence, the characteristic cargo quantity sequence only includes the cargo quantity corresponding to each characteristic date in the past period of time (such as holidays, dates of special events), thereby obtaining the non-characteristic cargo quantity sequence after excluding the characteristic cargo quantity sequence; then, according to the characteristic cargo quantity sequence and the non-characteristic cargo quantity sequence, generate a characteristic effect cargo quantity sequence and a periodic cargo quantity sequence, and then subtract the periodic cargo quantity sequence from the non-characteristic cargo quantity sequence to obtain the historical conventional cargo quantity sequence that is not affected by factors such as holidays and cycles.

[0104] Optionally, a characteristic effect sequence and a periodic cargo volume sequence are generated according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence, including: performing linear interpolation processing on the non-characteristic cargo volume sequence to obtain an interpolated cargo volume sequence; subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence to obtain a characteristic effect sequence; decomposing the non-characteristic cargo volume sequence according to different periodic dimensions and using a time series decomposition algorithm to obtain sub-period cargo volume sequences corresponding to each periodic dimension; generating a periodic cargo volume sequence according to the sub-period cargo volume sequences corresponding to each periodic dimension.

[0105] The first step of sequence decomposition is to decompose the characteristic cargo volume sequence affected by factors such as holidays and dates of special events. Specifically, based on business experience and data analysis, cargo volume is affected by different holidays and special events. For example, it is lower during the Mid-Autumn Festival and National Day, while it is higher during the Double 11 promotion. It is necessary to first mark the holidays and special dates in the past and future. The following is an example of marking the characteristic dates that affect cargo volume in 2023.

[0106]

[0107] According to the characteristic dates marked in the original cargo volume sequence, the characteristic cargo volume sequence is extracted from the original cargo volume sequence, and then the non-characteristic cargo volume sequence from which the characteristic cargo volume sequence has been eliminated is subjected to linear interpolation (according to the start time and end time of the characteristic date, the cargo volume records of the two days with the start time and end time in the non-characteristic cargo volume sequence are connected and interpolated for each day in between), thus obtaining the interpolated cargo volume sequence. Then, the characteristic effect sequence can be obtained by subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence.

[0108] The second step of sequence decomposition is the decomposition of the periodic effect. Specifically, according to different periodic dimensions, the non-characteristic cargo volume sequence after removing the characteristic cargo volume sequence is decomposed using the time series decomposition algorithm to obtain the sub-period cargo volume sequence corresponding to each periodic dimension. Optionally, the non-characteristic cargo volume sequence is first aggregated by month, and then the time series decomposition algorithm (STL, Seasonal and Trend decomposition using Loess) is used with a period of 12 to decompose the non-characteristic cargo volume sequence aggregated by month into an annual sub-period cargo volume sequence. Then, in the daily dimension, the time series decomposition algorithm (STL) is used with periods of 30 and 7 respectively to decompose it into a monthly sub-period cargo volume sequence and a weekly sub-period cargo volume sequence in the length of one year. After decomposing these three sub-period cargo volume sequences, the sequence synthesis method is used to synthesize the annual, monthly, and weekly sub-period cargo volume sequences to obtain the final periodic cargo volume sequence.

[0109] The STL method decomposes time series data into three components: trend, seasonality, and residual. The core idea of ​​the STL method is to estimate the trend and seasonality components of time series data through local weighted regression (Loess). The advantage of the STL method is that it can consider both long-term trends and seasonal changes, and can handle nonlinear and non-periodic time series data.

[0110] Optionally, generating a periodic cargo volume sequence according to the sub-period cargo volume sequences corresponding to the various periodic dimensions includes: adding the sub-period cargo volume sequences corresponding to the various periodic dimensions to obtain a periodic cargo volume sequence; wherein the various periodic dimensions include a year dimension, a day dimension under a month dimension, and a day dimension under a week dimension. In an embodiment of the present invention, for each day in a future period of time, the sub-period cargo volume sequences corresponding to the various periodic dimensions are added to obtain a periodic cargo volume sequence in the future period of time.

[0111] After calculating the periodic cargo volume sequence, subtract the periodic cargo volume sequence from the non-characteristic cargo volume sequence to obtain the historical regular cargo volume sequence that is not affected by holidays, cycles and other factors. According to the marked characteristic dates in the future, add the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain the fluctuating cargo volume sequence affected by holidays, cycles and other factors in the future.

[0112] It should be noted that if there are no holidays or special events in the future, there is no need to add the characteristic effect volume sequence to the periodic volume sequence. In this case, the periodic volume sequence is the fluctuating volume sequence.

[0113] Step 102: The historical conventional cargo volume sequence is used to train a cargo volume prediction model corresponding to the transport route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in the future.

[0114] In this step, for each transport route, the historical regular cargo volume sequence corresponding to the transport route that is not affected by holidays, cycles and other factors is used to train the cargo volume prediction model corresponding to the transport route, and then the cargo volume prediction model is used to predict the predicted regular cargo volume sequence of the transport route in the future (such as the next week, the next two weeks, the next month, the next three months, the next six months or the next year, etc.).

[0115] Step 103: Generate a predicted cargo volume sequence for the transport route in a future period of time based on the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence.

[0116] Since the historical regular cargo volume sequence is extracted from the original cargo volume sequence in step 101, the fluctuating cargo volume sequence affected by holidays, cycles, etc. in the future can be further extracted from the original cargo volume sequence. After predicting the predicted regular cargo volume sequence in the future, the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence are added together to obtain the predicted cargo volume sequence of the transport route in the future.

[0117] Optionally, step 103 may include: adding the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain a fluctuating cargo volume sequence; extracting the fluctuating cargo volume sequence corresponding to the predicted conventional cargo volume sequence from the fluctuating cargo volume sequence; adding the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence to obtain a predicted cargo volume sequence for the transport route in the future. The fluctuating cargo volume sequence may be caused by holidays, special events, etc. Therefore, the embodiment of the present invention decomposes the original cargo volume sequence into a historical conventional cargo volume sequence and a fluctuating cargo volume sequence, and then extracts the fluctuating cargo volume sequence corresponding to the predicted conventional cargo volume sequence from the fluctuating cargo volume sequence, thereby accurately predicting the predicted cargo volume sequence for the transport route in the future.

[0118] It should be noted that the length of the future period of time may be the same as the length of the past period of time, or may be shorter than the length of the past period of time, and the embodiment of the present invention does not limit this. If the length of the future period of time is the same as the length of the past period of time, the fluctuating cargo volume sequence is directly used as the fluctuating cargo volume sequence corresponding to the predicted conventional cargo volume sequence. If the length of the future period of time is shorter than the length of the past period of time, it is necessary to extract the fluctuating cargo volume sequence corresponding to the predicted conventional cargo volume sequence from the fluctuating cargo volume sequence based on the length of the future period of time.

[0119] Step 104, generating a vehicle dispatching plan based on the predicted cargo volume sequence of each route.

[0120] After accurately predicting the predicted cargo volume sequence of each route, a vehicle dispatching plan with the greatest reduction in transportation cost is generated based on the transportation cost of the first vehicle (such as a self-owned vehicle) and the transportation cost of the second vehicle (such as a third-party vehicle). In other words, the demand for future transportation tasks is analyzed, and tasks with stable demand, which can form a loop and have a high cost for third-party vehicles are found. Self-owned vehicles are allowed to replace third-party vehicles on these fixed loops, and transport as much as possible at full capacity to reduce the overall transportation cost.

[0121] According to the various embodiments described above, it can be seen that the embodiments of the present invention extract the historical conventional cargo volume sequence from the original cargo volume sequence, use the historical conventional cargo volume sequence to train to obtain the cargo volume prediction model corresponding to the transportation route, and then use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transportation route in the future. According to the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence, the technical means of generating the predicted cargo volume sequence of the transportation route in the future solves the technical problem of low vehicle loading rate leading to waste of transportation capacity and high transportation cost in the prior art. The embodiments of the present invention can accurately predict the cargo volume of each transportation route in the future, and generate the vehicle scheduling plan with the lowest transportation cost based on the predicted cargo volume, which not only improves the loading rate, but also reduces the transportation cost.

[0122] Figure 2 is a flow chart of a method for generating a vehicle dispatching plan according to a reference embodiment of the present invention. As another embodiment of the present invention, Figure 2 As shown, the method for generating a vehicle dispatching plan may include:

[0123] Step 201, obtaining the original cargo volume sequence of the transport route in the past period of time, and decomposing the original cargo volume sequence into a historical regular cargo volume sequence and a fluctuating cargo volume sequence.

[0124] First, obtain the daily waybill data of each transport route in the past period of time (such as the past year, the past two years, the past three years or the past five years, etc.), and then aggregate them by day to obtain the daily historical cargo volume of each transport route, and then fill the missing data of the aggregated historical cargo volume to obtain the original cargo volume sequence of each transport route, and finally decompose the original cargo volume sequence of each transport route into the historical regular cargo volume sequence and the fluctuating cargo volume sequence of the route. The embodiment of the present invention decomposes the original cargo volume sequence into the historical regular cargo volume sequence and the fluctuating cargo volume sequence, and uses the historical regular cargo volume sequence to train the cargo volume prediction model, which can improve the prediction accuracy of the model.

[0125] Step 202, determining each historical duration according to a preset duration interval and a preset step length.

[0126] For example, the preset duration interval is [30,180], the preset step length is 5 days, and the historical periods of 30 days, 35 days, 40 days, 45 days, 50 days, ..., 180 days are determined. For example, the preset duration interval is [20,150], the preset step length is 10 days, and the historical periods of 20 days, 30 days, 40 days, 50 days, 60 days, ..., 150 days are determined. For example, the preset duration interval is [10,270], the preset step length is 20 days, and the historical periods of 10 days, 30 days, 50 days, 70 days, 90 days, ..., 270 days are determined. The duration interval and step length can be set according to actual needs, and the embodiments of the present invention do not limit this.

[0127] Step 203: for each historical duration, a historical conventional cargo volume sequence corresponding to the historical duration is used to train a cargo volume prediction model corresponding to the transportation route, and the average absolute percentage error of the cargo volume prediction model is calculated.

[0128] For each historical duration, the historical conventional cargo volume sequence corresponding to the historical duration is used to train the cargo volume prediction model corresponding to the transport route, and then the average absolute percentage error of the cargo volume prediction model is calculated. Optionally, the cargo volume prediction model can be a normal distribution model, a log-normal distribution model, or an XGBoost model, which is not limited in the embodiment of the present invention.

[0129] Optionally, a cargo volume prediction model corresponding to the transport route is obtained by training using a historical conventional cargo volume sequence corresponding to the historical duration, and an average absolute percentage error of the cargo volume prediction model is calculated, including: fitting a log-normal distribution model using a historical conventional cargo volume sequence corresponding to the historical duration, thereby training a cargo volume prediction model corresponding to the transport route; using the cargo volume prediction model to predict a predicted cargo volume sequence for a future period of time corresponding to the historical duration; and calculating the average absolute percentage error of the cargo volume prediction model based on the predicted cargo volume sequence for the future period of time corresponding to the historical duration and the actual cargo volume sequence for the future period of time corresponding to the historical duration.

[0130] The cargo volume prediction model of the embodiment of the present invention is slightly different from the general prediction model, because in the transport coordination scenario, the cargo volume prediction tends to be as close to the actual cargo volume as possible while being less than the actual cargo volume. Based on this feature, the embodiment of the present invention specifically designs a cargo volume prediction method. Through data analysis, it is found that the cargo volume of the route obeys the lognormal distribution, and at the same time, the random variable of the cargo volume of the route obeys the weak Markov hypothesis, that is, the route has a stronger correlation with recent history. Based on the distribution and hypothesis, a cargo volume prediction model is designed. This model is based on the lognormal distribution, and the prediction method is to output the probability quantile of this random variable of 5%, 10%, 12%, 20%, 25% or 30% (which can be adjusted according to actual needs). There are two parameters in this model (mean and standard deviation) that need to be fitted from historical data.

[0131] For each historical duration, the log-normal distribution model is fitted using the historical conventional cargo volume sequence corresponding to the historical duration, thereby training the cargo volume forecasting model corresponding to the transport route. When evaluating the forecasting effect, the cargo volume for a period of time in the future is rolled out at a preset frequency (such as one week, one month, or three months) within the past period of time, and then the average absolute percentage error of the cargo volume forecasting model is calculated using the predicted cargo volume sequence and the actual cargo volume sequence.

[0132] Optionally, the average absolute percentage error of the cargo volume prediction model is calculated based on the predicted cargo volume sequence in the future period corresponding to the historical duration and the actual cargo volume sequence in the future period corresponding to the historical duration, including: for each time point in the future period corresponding to the historical duration, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point; if so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error; if not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume; calculate the average absolute percentage error of the cargo volume prediction model based on the absolute percentage errors corresponding to each time point and / or the absolute percentage errors after reducing the penalty. For each time point in the future period corresponding to the historical time length, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point. If so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error, such as removing the logarithm of the absolute percentage error or dividing the absolute percentage error by an integer greater than one (such as 2, 3 or 4, etc.), which can further reduce the penalty for the smaller prediction. If not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume. Finally, calculate the average absolute percentage error of the cargo volume forecasting model based on the absolute percentage error corresponding to each time point and / or the absolute percentage error after reducing the penalty.

[0133] Step 204, screen out the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration.

[0134] For this route, after calculating the average absolute percentage error corresponding to each cargo volume prediction model, the cargo volume prediction model with the smallest average absolute percentage error is screened out, and then the historical duration corresponding to the cargo volume prediction model is determined, thereby obtaining the final training result.

[0135] In terms of evaluation indicators, the prior art uses MAPE (mean absolute percentage error) for evaluation, but the embodiment of the present invention improves the evaluation indicator by reducing the penalty for absolute percentage error when the predicted cargo volume is less than or equal to the actual cargo volume, that is, asymmetric-MAPE.

[0136] mape (Mean Absolute Percentage Error) is the mean absolute percentage error, which is a commonly used indicator to measure the accuracy of a prediction model or method in statistics. mape represents the average value of the absolute value of the difference between the predicted value and the actual value, expressed as a percentage. asymmetric-mape represents asymmetric mape (this evaluation index is more suitable for the prediction evaluation of the business scenario of the present invention, and is more suitable for the transportation scheduling scenario), that is, on the basis of mape, it penalizes the part of the prediction that is too high. 1-mape and 1-asymmetric-mape are processed for intuitiveness, and the larger the result minus the evaluation index, the better.

[0137] The results of the two evaluation indicators are compared with the prediction model of the prior art and the prediction model of the embodiment of the present invention, as shown below:

[0138]

[0139] By comparison, it can be seen that no matter which evaluation index is used, the prediction effect of the present invention is significantly better than that of the prior art.

[0140] Step 205, using the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration, as well as a preset probability quantile, to forecast the predicted regular cargo volume sequence of the transportation route in a future period of time.

[0141] In this step, the cargo volume forecasting model with the smallest average absolute percentage error selected in step 204 and its corresponding historical duration, as well as a preset probability quantile (such as 5%, 10%, 12%, 20%, 25% or 30%) are used to predict the predicted regular cargo volume sequence of the transport route in a period of time in the future (such as the next week, the next month, the next three months or the next six months, etc.).

[0142] It should be noted that if you want the loading rate of the first vehicle (such as a self-owned vehicle) to be as high as possible, you can set the probability quantile to 5%. If you want the goods to be delivered as soon as possible, you can accept a lower vehicle loading rate and set the probability quantile to 50% (even greater than 50%, but it is not recommended to do so. This parameter is too large and may make it impossible to reduce costs). For example, if the probability quantile is set to 20%, it means that 80% of the predicted cargo volume is less than the actual cargo volume. For example, if the probability quantile is set to 5%, it means that 95% of the predicted cargo volume is less than the actual cargo volume. Optionally, the embodiment of the present invention sets the probability quantile to any value between 15-25%.

[0143] Step 206: Generate a predicted cargo volume sequence for the transport route in a future period of time based on the predicted regular cargo volume sequence and the fluctuating cargo volume sequence.

[0144] The predicted regular cargo volume series and the fluctuating cargo volume series are added together to obtain the predicted cargo volume series for the transport route in the future.

[0145] The method of the embodiment of the present invention can more accurately predict stable cargo volume and fluctuating cargo volume, so that the final predicted cargo volume meets the needs of overall transportation scheduling, which helps to improve the loading rate and reduce transportation costs.

[0146] like Figure 3 As shown, taking a certain route from September 15th to October 15th as an example, the actual cargo volume, the cargo volume predicted by the existing solution, and the cargo volume predicted by the present invention are compared. The existing technical solution is based on the historical shipment model exceeding the threshold. The total volume of the models actually used on the general route is greater than or equal to the daily total cargo volume. The prediction based on this usually results in the predicted cargo volume being greater than the actual cargo volume. The prediction method of the present invention takes the low probability quantile after historical statistical learning, so the overall prediction result is mostly less than or equal to and close to the actual cargo volume. Moreover, for dates corresponding to holidays and special events, such as Mid-Autumn Festival and National Day, as well as ordinary week fluctuations, the current technical solution is unable to predict, but the embodiment of the present invention can refer to the history of previous years to predict certain fluctuations.

[0147] Step 207, generating a vehicle dispatching plan based on the predicted cargo volume sequence of each route.

[0148] After accurately predicting the predicted cargo volume sequence of each route, a vehicle scheduling plan with the greatest reduction in transportation cost is generated based on the transportation cost of the first vehicle (such as a self-owned vehicle) and the transportation cost of the second vehicle (such as a third-party vehicle) and based on a greedy algorithm.

[0149] The embodiment of the present invention can accurately predict the cargo volume of each transport route in the future, and generate a vehicle dispatching plan with the lowest transportation cost based on the predicted cargo volume, which not only improves the loading rate but also reduces the transportation cost. Therefore, the embodiment of the present invention solves the technical problem of low loading rate and high transportation cost caused by inaccurate cargo volume prediction in existing transportation scheduling. Moreover, the embodiment of the present invention introduces the probability quantile of cargo volume prediction, and can adjust the probability quantile to control the plan to obtain the best adaptation plan in terms of different indicators such as loading rate, timeliness, and cost reduction.

[0150] In addition, the specific implementation content of the method for generating a vehicle dispatching plan in a reference embodiment of the present invention has been described in detail in the method for generating a vehicle dispatching plan described above, so the repeated content will not be described here.

[0151] Figure 4 , which is a flow chart of a method for generating a vehicle dispatching scheme according to another reference embodiment of the present invention. Figure 4 As shown, the method for generating a vehicle dispatching plan may include:

[0152] Step 401, obtaining the original cargo volume sequence of the transport route in the past period of time, and extracting the historical regular cargo volume sequence from the original cargo volume sequence.

[0153] Step 402: Use the historical conventional cargo volume sequence to train and obtain a cargo volume prediction model corresponding to the transportation route.

[0154] Step 403: Use the cargo volume prediction model to predict the predicted regular cargo volume sequence of the transportation route in the future.

[0155] Step 404 , the predicted regular cargo volume sequence is added to the fluctuating cargo volume sequence corresponding to the predicted regular cargo volume sequence to generate a predicted cargo volume sequence for the transport route in a future period of time.

[0156] Step 405 , matching each route with a plurality of first vehicles and a plurality of second vehicles one by one according to the predicted cargo volume sequence of each route.

[0157] Step 406: Based on the transportation cost of the first vehicle and the transportation cost of the second vehicle, a vehicle dispatching solution that minimizes the total transportation cost is solved using a greedy algorithm.

[0158] After predicting the cargo volume of each transportation route in the future, the initial position parameters and transportation costs of each model of the first vehicle (self-owned vehicle) and the transportation costs of each model of the second vehicle (third-party vehicle) are obtained, and then the greedy algorithm is used to explore the scheduling solution. The greedy algorithm is a heuristic algorithm that selects the best decision under the current state at each step in the hope of eventually reaching the global optimal solution. The greedy algorithm can be used to solve the vehicle scheduling solution with the greatest reduction in total transportation cost.

[0159] The following are the main steps to generate a vehicle scheduling solution using a greedy algorithm:

[0160] Initialization: All unassigned route cargo and available vehicles are placed in the candidate pool, and the scheduling plan is initialized. The optimization goal is defined as maximizing transportation cost reduction, and the average of the transportation cost quotations of third-party vehicles on the same route in the past 30 days is used as the transportation cost quotation of third-party vehicles. The transportation cost accounting of self-operated vehicles follows the existing cost accounting logic.

[0161] Vehicle selection: Choose the largest vehicle among the available vehicles and located at the starting point of the route with the largest cargo volume, taking into account factors such as the vehicle's release time, load capacity and loading time.

[0162] Loop line task generation: Refer to the departure and arrival times of the route's historical 30-day shipping tasks to generate transportation tasks for the predicted cargo volume. When the predicted cargo volume is greater than the shipping task, a greedy algorithm is used to generate transportation tasks. After generating a task for each route, prioritize searching for transportation tasks that can form a single-vehicle loop. Since there are a few days in the middle when the predicted cargo volume is low and the generated tasks have smaller vehicle models, there are a few small vehicle models in the 30-day loop task. The second search is to search for loop line combinations that consider other sufficient cargo volumes on these dates with smaller cargo volumes.

[0163] Vehicle task matching: After completing all loop task mining, select the task combination with the best cost reduction in the next 30 days for the selected vehicle results.

[0164] Scheduling update: Assign the selected tasks to the selected vehicles and update the scheduling plan and the alternative pool.

[0165] Termination condition: Repeat the steps of vehicle selection, loop task mining, and vehicle task matching until all vehicles are performing tasks or the cost cannot be further reduced by performing tasks.

[0166] In the process of greedy algorithm, factors such as transportation cost, time window, vehicle utilization rate, etc. need to be comprehensively considered to generate the optimal overall scheduling plan. Appropriate evaluation function or objective function can be defined to make trade-offs and optimizations according to specific needs. By reasonably arranging transportation tasks and optimizing the allocation of vehicles, operating costs can be reduced, transportation efficiency can be improved, and tasks can be completed within the time window.

[0167] After the vehicle dispatch plan is generated, the cost reduction amount of the plan will be calculated, the time window for the arrival of the goods (for example, in plan A, 95% of the goods will arrive 1 hour earlier than planned, and in plan B, the goods are expected to arrive at the same time as planned), and the vehicle utilization rate (for example, in plan A, the utilization rate of self-owned vehicles is 85%, and in plan B, the utilization rate is 80%). The greedy algorithm will only find the plan with the greatest cost reduction, and the introduction of these indicators is to consider listing the top cost reduction plans, so that operators can comprehensively consider the above factors.

[0168] Since the current cargo volume forecast has a predicted value and a probability distribution, the probability threshold can be adjusted based on business needs to adjust the cargo volume forecast. For example, the probability quantile is set to 20% (the probability quantile here means that in the backtest history, the predicted cargo volume will be greater than the actual cargo volume in 20% of cases, which may cause the volume of the dispatched vehicle to be greater than the transportation demand). After selecting the probability quantile, the cargo volume forecast model outputs the cargo volume forecast value. The actual optimized dispatched vehicle model may be less than the actual total cargo volume of the route (the excess cargo volume is further dispatched or a third-party small vehicle is used as a backup), so the probability quantile can be slightly increased. Raising this parameter will improve transportation timeliness and slightly reduce transportation loading rate and vehicle efficiency. Specific adjustments can be made by backtesting the history, based on the needs of the operation scenario, and taking into account timeliness, loading rate, cost reduction, etc. The optional space for the probability quantile is 5%-50%.

[0169] If the existing technology method is used, there are about 500 transportation tasks in a certain area from September 15 to October 15, the average loading rate is 0.72, and the proportion of low loading (main line loading rate is less than 50%, branch line loading rate is less than 40%) is 0.06. However, using the method of the embodiment of the present invention, the predicted probability quantile is set to 20%, and the vehicle scheduling plan is output. The average loading rate is 0.76, the low loading ratio is 0.03, the backtest timeliness index has no obvious difference, and the loading rate is significantly improved.

[0170] In addition, the specific implementation content of the method for generating a vehicle dispatching plan in another reference embodiment of the present invention has been described in detail in the method for generating a vehicle dispatching plan described above, so the repeated content will not be described here.

[0171] Figure 5 Schematic diagram of a device for generating a vehicle dispatching plan according to an embodiment of the present invention. Figure 5 As shown, the device 500 for generating a vehicle scheduling plan includes a first sequence module 501, a prediction module 502, a second sequence module 503 and a generation module 504; wherein the first sequence module 501 is used to obtain the original cargo volume sequence of the transportation route in the past period of time, and extract the historical conventional cargo volume sequence from the original cargo volume sequence; the prediction module 502 is used to use the historical conventional cargo volume sequence to train to obtain a cargo volume prediction model corresponding to the transportation route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transportation route in the future period of time; the second sequence module 503 is used to generate a predicted cargo volume sequence of the transportation route in the future period of time according to the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence; the generation module 504 is used to generate a vehicle scheduling plan according to the predicted cargo volume sequence of each route.

[0172] Optionally, the first sequence module 501 is further used for:

[0173] Extracting a characteristic cargo quantity sequence from the original cargo quantity sequence, thereby obtaining a non-characteristic cargo quantity sequence after removing the characteristic cargo quantity sequence;

[0174] Generate a characteristic effect cargo volume sequence and a periodic cargo volume sequence according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence;

[0175] The periodic cargo volume sequence is subtracted from the non-characteristic cargo volume sequence to obtain the historical regular cargo volume sequence.

[0176] Optionally, the first sequence module 501 is further used for:

[0177] Performing linear interpolation processing on the non-characteristic cargo quantity sequence to obtain an interpolated cargo quantity sequence;

[0178] Subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence to obtain a characteristic effect sequence;

[0179] Decomposing the non-characteristic cargo volume sequence according to different period dimensions and using a time series decomposition algorithm, thereby obtaining a sub-period cargo volume sequence corresponding to each period dimension;

[0180] Generate a periodic cargo quantity sequence according to the sub-period cargo quantity sequences corresponding to each periodic dimension.

[0181] Optionally, the first sequence module 501 is further used for:

[0182] Add the sub-period cargo volume sequences corresponding to each period dimension to obtain a period cargo volume sequence;

[0183] The various periodic dimensions include the year dimension, the day dimension under the month dimension, and the day dimension under the week dimension.

[0184] Optionally, the prediction module 502 is further configured to:

[0185] Determine each historical duration according to a preset duration interval and a preset step length;

[0186] For each historical duration, a cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence corresponding to the historical duration, and the average absolute percentage error of the cargo volume prediction model is calculated;

[0187] Filter out the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration;

[0188] The cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration, as well as the preset probability quantile, is used to forecast the predicted regular cargo volume sequence of the transport route in the future.

[0189] Optionally, the prediction module 502 is further configured to:

[0190] The log-normal distribution model is fitted using the historical conventional cargo volume sequence corresponding to the historical duration, so as to train and obtain a cargo volume prediction model corresponding to the transportation route;

[0191] Using the cargo volume forecasting model to forecast a forecast cargo volume sequence for a future period of time corresponding to the historical time length;

[0192] The average absolute percentage error of the cargo volume prediction model is calculated based on the predicted cargo volume sequence in the future period corresponding to the historical time length and the actual cargo volume sequence in the future period corresponding to the historical time length.

[0193] Optionally, the prediction module 502 is further configured to:

[0194] For each time point in the future period corresponding to the historical time length, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point; if so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error; if not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume;

[0195] The average absolute percentage error of the cargo volume prediction model is calculated according to the absolute percentage errors corresponding to the respective time points and / or the absolute percentage errors after reducing the penalty.

[0196] Optionally, the second sequence module 503 is further used for:

[0197] Add the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain a fluctuating cargo volume sequence;

[0198] Extracting a fluctuating cargo volume sequence corresponding to the predicted regular cargo volume sequence from the fluctuating cargo volume sequence;

[0199] The predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence are added together to obtain the predicted cargo volume sequence of the transport route in a future period of time.

[0200] Optionally, the generating module 504 is further configured to:

[0201] According to the predicted cargo volume sequence of each route, each route is matched with a plurality of first vehicles and a plurality of second vehicles one by one;

[0202] According to the transportation cost of the first vehicle and the transportation cost of the second vehicle, a vehicle dispatching scheme with the greatest reduction in total transportation cost is solved by using a greedy algorithm.

[0203] It should be noted that the specific implementation content of the device for generating a vehicle dispatching plan in the present invention has been described in detail in the method for generating a vehicle dispatching plan described above, so the repeated content will not be described here.

[0204] Figure 6 An exemplary system architecture 600 is shown to which the method for generating a vehicle dispatch plan or the apparatus for generating a vehicle dispatch plan according to the embodiments of the present invention can be applied.

[0205] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0206] The user can use the terminal devices 601, 602, 603 to interact with the server 605 through the network 604 to receive or send messages, etc. Various communication client applications can be installed on the terminal devices 601, 602, 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0207] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0208] The server 605 may be a server that provides various services, such as a backend management server (for example only) that provides support for shopping websites browsed by users using the terminal devices 601, 602, and 603. The backend management server may analyze and process the received data such as the item information query request, and feed back the processing results to the terminal device.

[0209] It should be noted that the method for generating a vehicle dispatching scheme provided in the embodiment of the present invention is generally executed by the server 605, and accordingly, the device for generating a vehicle dispatching scheme is generally set in the server 605. The method for generating a vehicle dispatching scheme provided in the embodiment of the present invention can also be executed by the terminal devices 601, 602, and 603, and accordingly, the device for generating a vehicle dispatching scheme can be set in the terminal devices 601, 602, and 603.

[0210] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0211] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system 700 of a terminal device suitable for implementing an embodiment of the present invention. Figure 7 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0212] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 to a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0213] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0214] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0215] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0216] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0217] The modules involved in the embodiments of the present invention may be implemented by software or hardware. The modules described may also be set in a processor, for example, it may be described as: a processor includes a first sequence module, a prediction module, a second sequence module and a generation module, wherein the names of these modules do not constitute a limitation on the modules themselves in some cases.

[0218] As another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by a device, the device implements the following method: obtaining the original cargo volume sequence of the transportation route in the past period of time, extracting the historical conventional cargo volume sequence from the original cargo volume sequence; using the historical conventional cargo volume sequence to train and obtain the cargo volume prediction model corresponding to the transportation route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transportation route in the future period of time; based on the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence, generating the predicted cargo volume sequence of the transportation route in the future period of time; generating a vehicle scheduling plan based on the predicted cargo volume sequence of each route.

[0219] As another aspect, an embodiment of the present invention further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method described in any of the above embodiments is implemented.

[0220] According to the technical solution of the embodiment of the present invention, because the historical conventional cargo volume sequence is extracted from the original cargo volume sequence, the cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence, and the predicted conventional cargo volume sequence of the transportation route in the future period is predicted by the cargo volume prediction model, and the predicted conventional cargo volume sequence and its corresponding fluctuating cargo volume sequence are used to generate the predicted cargo volume sequence of the transportation route in the future period. Therefore, the technical problem of low vehicle loading rate leading to waste of transportation capacity and high transportation cost in the prior art is overcome. The embodiment of the present invention can accurately predict the cargo volume of each transportation route in the future period, and generate the vehicle scheduling plan with the lowest transportation cost based on the predicted cargo volume, which not only improves the loading rate but also reduces the transportation cost.

[0221] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for generating a vehicle dispatching plan, characterized in that: include: Obtaining an original cargo volume sequence of a transport route in the past period of time, and extracting a historical regular cargo volume sequence from the original cargo volume sequence; The historical conventional cargo volume sequence is used to train a cargo volume prediction model corresponding to the transport route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in the future; Generate a predicted cargo volume sequence for the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence; Generate a vehicle dispatch plan based on the predicted cargo volume sequence of each route.

2. The method according to claim 1, characterized in that: A historical conventional cargo volume sequence is extracted from the original cargo volume sequence, including: Extracting a characteristic cargo quantity sequence from the original cargo quantity sequence, thereby obtaining a non-characteristic cargo quantity sequence after removing the characteristic cargo quantity sequence; Generate a characteristic effect cargo volume sequence and a periodic cargo volume sequence according to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence; The periodic cargo volume sequence is subtracted from the non-characteristic cargo volume sequence to obtain the historical regular cargo volume sequence.

3. The method according to claim 2, characterized in that According to the characteristic cargo volume sequence and the non-characteristic cargo volume sequence, a characteristic effect sequence and a periodic cargo volume sequence are generated, including: Performing linear interpolation processing on the non-characteristic cargo quantity sequence to obtain an interpolated cargo quantity sequence; Subtracting the interpolated cargo volume sequence from the characteristic cargo volume sequence to obtain a characteristic effect sequence; Decomposing the non-characteristic cargo volume sequence according to different period dimensions and using a time series decomposition algorithm, thereby obtaining a sub-period cargo volume sequence corresponding to each period dimension; Generate a periodic cargo quantity sequence according to the sub-period cargo quantity sequences corresponding to each periodic dimension.

4. The method according to claim 3, characterized in that Generate a periodic cargo quantity sequence according to the sub-period cargo quantity sequence corresponding to each periodic dimension, including: Add the sub-period cargo volume sequences corresponding to each period dimension to obtain a period cargo volume sequence; The various periodic dimensions include the year dimension, the day dimension under the month dimension, and the day dimension under the week dimension.

5. The method according to claim 1, characterized in that The cargo volume prediction model corresponding to the transport route is obtained by training the historical conventional cargo volume sequence, so as to use the cargo volume prediction model to predict the conventional cargo volume sequence of the transport route in the future, including: Determine each historical duration according to a preset duration interval and a preset step length; For each historical duration, a cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence corresponding to the historical duration, and the average absolute percentage error of the cargo volume prediction model is calculated; Filter out the cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration; The cargo volume forecasting model with the smallest mean absolute percentage error and its corresponding historical duration, as well as the preset probability quantile, is used to forecast the predicted regular cargo volume sequence of the transport route in the future.

6. The method according to claim 5, characterized in that The cargo volume prediction model corresponding to the transportation route is obtained by training the historical conventional cargo volume sequence corresponding to the historical duration, and the average absolute percentage error of the cargo volume prediction model is calculated, including: The log-normal distribution model is fitted using the historical conventional cargo volume sequence corresponding to the historical duration, so as to train and obtain a cargo volume prediction model corresponding to the transportation route; Using the cargo volume forecasting model to forecast a forecast cargo volume sequence for a future period of time corresponding to the historical time length; The average absolute percentage error of the cargo volume prediction model is calculated based on the predicted cargo volume sequence in the future period corresponding to the historical time length and the actual cargo volume sequence in the future period corresponding to the historical time length.

7. The method according to claim 6, characterized in that Calculating the mean absolute percentage error of the cargo volume prediction model according to the predicted cargo volume sequence in the future period corresponding to the historical duration and the actual cargo volume sequence in the future period corresponding to the historical duration, including: For each time point in the future period corresponding to the historical time length, determine whether the predicted cargo volume corresponding to the time point is less than or equal to the actual cargo volume corresponding to the time point; if so, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume, and reduce the penalty for the absolute percentage error; if not, calculate the absolute percentage error between the predicted cargo volume and the actual cargo volume; The average absolute percentage error of the cargo volume prediction model is calculated according to the absolute percentage errors corresponding to the respective time points and / or the absolute percentage errors after reducing the penalty.

8. The method according to claim 2, characterized in that: Generate a predicted cargo volume sequence for the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence, including: Add the characteristic effect cargo volume sequence to the periodic cargo volume sequence to obtain a fluctuating cargo volume sequence; Extracting a fluctuating cargo volume sequence corresponding to the predicted regular cargo volume sequence from the fluctuating cargo volume sequence; The predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence are added together to obtain the predicted cargo volume sequence of the transport route in a future period of time.

9. The method according to claim 1, characterized in that: Generate vehicle dispatching plan based on the predicted cargo volume sequence of each route, including: According to the predicted cargo volume sequence of each route, each route is matched with a plurality of first vehicles and a plurality of second vehicles one by one; According to the transportation cost of the first vehicle and the transportation cost of the second vehicle, a vehicle dispatching scheme with the greatest reduction in total transportation cost is solved by using a greedy algorithm.

10. A device for generating a vehicle dispatching plan, characterized in that: include: The first sequence module is used to obtain the original cargo volume sequence of the transportation route in the past period of time, and extract the historical regular cargo volume sequence from the original cargo volume sequence; A prediction module, used to train the historical conventional cargo volume sequence to obtain a cargo volume prediction model corresponding to the transport route, so as to use the cargo volume prediction model to predict the predicted conventional cargo volume sequence of the transport route in the future; A second sequence module is used to generate a predicted cargo volume sequence for the transport route in a future period of time according to the predicted regular cargo volume sequence and its corresponding fluctuating cargo volume sequence; The generation module is used to generate a vehicle dispatching plan based on the predicted cargo volume sequence of each route.

11. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 9.

12. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.