Container wharf gate short-term container collection quantity prediction method and system and storage medium
By decomposing the route feature of the historical data of the container terminal and integrating multiple timing prediction algorithms, the problem of low prediction accuracy in the existing technology is solved, and high-precision prediction of the container collection volume in the next 24 hours is achieved, which improves the efficiency and flexibility of dock resource allocation.
Patent Information
- Application Number
- CN202510111339.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-23
- Publication Date
- 2025-06-03
AI Technical Summary
The existing container terminal gate short-term container collection prediction method is not high when dealing with complex route characteristics and variable container volume fluctuations, and it is difficult to cope with nonlinear changes and variable data.
By obtaining the historical box collection data of the container terminal, the box collection task is decomposed into the box collection time series of each route based on the route characteristics, and independent prediction is made using various timing prediction algorithms such as ARIMA, PROPHET and LSTM. Finally, the prediction results are integrated through the weighted average method to output the total box collection in the next 24 hours.
It improves the accuracy of short-term box collection prediction, can more accurately capture the unique patterns and complex changes of each route, and enhances the flexibility of terminal resource allocation and prediction stability.
Smart Images
Figure CN120087785A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of container terminal management, and in particular to a method, system and storage medium for predicting the short-term incoming container volume at the container terminal gate. Background Art
[0002] Container terminals play an important role in global logistics, and the incoming container operation at the gate is crucial for the overall terminal efficiency. Accurately predicting the incoming container volume in the short term helps the terminal to allocate resources reasonably, reduce congestion and improve operation efficiency. However, the existing methods for predicting the incoming container volume have many deficiencies in practical applications, resulting in low prediction accuracy and difficulty in coping with complex route characteristics and changing container volume fluctuations.
[0003] The disadvantages of the existing methods include:
[0004] Lack of independent analysis of route characteristics: Most traditional prediction methods model the total incoming container volume as a whole without decomposing the characteristics of different routes. This method is difficult to capture the unique patterns of each route, resulting in large deviations in the prediction results.
[0005] Insufficient handling of randomness and non-linear changes: The existing methods mainly rely on linear or simple statistical models and cannot effectively handle non-linear trends, seasonal fluctuations or long-term dependencies in time series. This limitation makes the model difficult to cope with the complex changes of container volume, especially in the face of holiday effects, voyage fluctuations, etc., and the prediction results are not accurate enough.
[0006] Lack of integration and optimization of multiple algorithms: The existing methods usually use a single algorithm for prediction and are difficult to take into account the respective advantages of different algorithms. For routes with linear changes, a simple linear model may have better results, while for routes with non-linear characteristics, deep learning algorithms will perform more excellently. The limitation of a single algorithm leads to insufficient generalization ability of the prediction model and difficulty in coping with different types of routes and data fluctuations. Summary of the Invention
[0007] The purpose of the present invention is to provide a method, system and storage medium for predicting the short-term incoming container volume at the container terminal gate, aiming to accurately predict the incoming container volume in the next 24 hours through the analysis of historical data and improve the efficiency and flexibility of terminal operations.
[0008] The present invention proposes a method for predicting the short-term incoming container volume at the container terminal gate, and the method includes:
[0009] Obtain the historical incoming container data of the container terminal, including routes, container volume and arrival time;
[0010] Preprocess the received container data, and decompose the container receiving tasks into the time series of container receiving volumes for each route based on route characteristics;
[0011] Use multiple time series prediction algorithms to independently predict the container receiving volumes for each route. The time series prediction algorithms include: ARIMA prediction algorithm, PROPHET prediction algorithm, and LSTM prediction algorithm;
[0012] Integrate the prediction results of the above three prediction algorithms for each route. The integration process includes:
[0013] Process outliers and eliminate extreme predicted values;
[0014] Align the prediction results of each route along the time axis;
[0015] Match corresponding weight values to the prediction results of the above three prediction algorithms through the weighted average method;
[0016] Output the prediction result of the total container receiving volume within the next 24 hours.
[0017] The method of the present invention is based on a decomposition and integration framework. By decomposing the total container receiving tasks by route, independently predicting the container volumes of each route, and integrating the prediction results to predict the container receiving demand in the short term in the future.
[0018] According to another aspect of the present invention, there is provided a short-term container receiving volume prediction system for a container terminal gate, including a data acquisition module, a data processing module, an algorithm prediction module, and a result integration module connected in sequence; wherein,
[0019] Data acquisition module: used to obtain the historical container receiving data of the container terminal, including route, container volume, and arrival time.
[0020] Data processing module: used to preprocess the collected data, and decompose the container receiving tasks into the time series of container receiving volumes for each route based on route characteristics.
[0021] Algorithm prediction module: respectively use the ARIMA algorithm, the PROPHET algorithm, and the LSTM algorithm to independently predict the container receiving volumes of each route:
[0022] Result integration module: integrate the prediction results of each route based on the methods of processing outliers, time axis alignment, and weighted average method, and output the total container receiving volume in the next 24 hours.
[0023] In the above technical solution, in order to better use the above method, the present application proposes a short-term container receiving volume prediction system for a container terminal gate. Each module corresponds to each step of the above method, and its specific principle has been described above and will not be elaborated here.
[0024] According to another aspect of the present invention, there is provided a computer-readable storage medium storing a computer program, which when executed by a processor implements the above method.
[0025] In the above technical solution, in order to better run and use the method, the above method is stored in a computer-readable storage medium, and the processor is used to implement the above method. It should be noted that the principle and effect of each step have been described above and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.
[0027] Figure 1 is a schematic flowchart of an embodiment of a method for predicting the short-term incoming container volume at the container terminal gate of the present invention;
[0028] Figure 2 is a schematic structural diagram of an embodiment of a system for predicting the short-term incoming container volume at the container terminal gate of the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS
[0029] The present invention will be further described in detail below with reference to the drawings and embodiments. It should be specifically noted that the following embodiments are only used to illustrate the present invention, but do not limit the scope of the present invention. Similarly, the following embodiments are only partial embodiments of the present invention rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.
[0030] The present invention provides a method, a system and a storage medium for predicting the short-term incoming container volume at the container terminal gate, which can generate a prediction result of the incoming container volume in a future period of time by combining route feature decomposition and various time series prediction algorithms. This method can solve the problem of insufficient accuracy of the traditional method for predicting the future incoming container volume, and reduce operation conflicts and resource waste by reasonably allocating terminal resources.
[0031] Embodiment 1
[0032] Please refer to Figure 1 , a method for predicting the short-term incoming container volume at the container terminal gate, the method comprising:
[0033] S1. Obtain the historical incoming container data of the container terminal, including the route, the container volume, and the arrival time;
[0034] In this embodiment, the container terminal's historical record system is used to obtain the container collection data in the recent period. The container collection data includes route classification, hourly container volume records, and arrival time, etc. The data range covers at least the past 12 months to capture the periodicity and fluctuation characteristics of the container volume and arrival time of each route.
[0035] S2, pre-processing the container collection data, and decomposing the container collection task into a time series of container collection volume of each route based on route characteristics;
[0036] In this embodiment, the pre-processing step of the received container data is to decompose the total container volume into the time series of the received container volume of each route based on the route classification, and the format is (t i ,y i ), where t i Indicates dividing route i into corresponding time points t, y i Indicates the number of boxes received at the corresponding time point t.
[0037] In the preprocessing process, the received box data needs to be cleaned and missing values supplemented. Among them, data cleaning means that there may be repeated records of received boxes in certain time periods, and these duplicate items can be cleaned up by deleting or merging. Missing value supplementation means that there may be missing records of received boxes in certain time periods, and these missing items can be filled with fixed values (such as mean, median or mode).
[0038] S3. Use multiple time series prediction algorithms to independently predict the box collection volume of each route. The time series prediction algorithms include: ARIMA prediction algorithm, PROPHET prediction algorithm and LSTM prediction algorithm;
[0039] In this embodiment, the ARIMA algorithm is suitable for routes with stable short-term trends, and performs differentiation, parameter fitting and prediction on time series with stable and obvious linear trends to improve the accuracy of prediction results.
[0040] Using PROPHET algorithm: It is suitable for routes with obvious fluctuations. For routes with periodic fluctuations, such as those strongly affected by seasonal weather and holiday effects, forecasts are made through trend decomposition and period fitting to improve the accuracy of forecast results.
[0041] Adopt LSTM algorithm: It is suitable for routes with complex fluctuations and strong long-term trends. It processes long-term dependencies and nonlinear changes through deep learning networks to improve the accuracy of prediction results.
[0042] S4. Integrate the prediction results of the above three prediction algorithms for each route. The integration process includes:
[0043] S41, process outliers and remove extreme prediction values;
[0044] The container receiving volume obtained through the ARIMA prediction algorithm, PROPHET prediction algorithm, and LSTM prediction algorithm can eliminate extreme prediction values through the median absolute deviation algorithm (MAD). This algorithm can measure the degree of dispersion of the container volume data of each route i relative to its central tendency (median) at a specific time point t. When it is observed that the deviation of the container receiving volume y i , from the median is greater than a certain multiple of MAD (such as 3.5 times), then this observed value can be considered an outlier, and this outlier is automatically eliminated and replaced with a moving average or median.
[0045] S42. Align the prediction results of each route on the time axis, and align the prediction results of different prediction algorithms according to the time range of the next 24 hours;
[0046] S43. Match the corresponding weight values to the prediction results of the above three prediction algorithms through the weighted average method;
[0047] In this embodiment, the weight value is based on the performance of each algorithm on the historical data set, and is determined by evaluating its mean absolute error or mean square error.
[0048] S5. Output the predicted result of the total container receiving volume within the next 24 hours to provide support for the terminal operation scheduling; the final predicted result of the total container receiving volume generates a time series curve, visually showing the hourly change trend of the total container receiving volume, and providing data support for the terminal resource allocation and task scheduling.
[0049] In this embodiment, the predicted result of the total container receiving volume within the next 24 hours is calculated by the following formula:
[0050]
[0051] In the formula: n represents the total number of routes, represents the predicted result calculated by the weighted average method
[0052] is calculated by the following formula:
[0053]
[0054] In the formula: represents the container receiving volume predicted by the ARIMA prediction algorithm for route i at time point t, represents the container receiving volume predicted by the PROPHET prediction algorithm for route i at time point t,
[0055] represents the container receiving volume predicted by the LSTM prediction algorithm for route i at time point t;
[0056] w ARIMA represents the prediction error of the ARIMA prediction algorithm;
[0057] w PROPHET represents the prediction error of the PROPHET prediction algorithm;
[0058] w LSTM represents the prediction error of the LSTM prediction algorithm.
[0059] The said w ARIMA 、w PROPHET 、w LSTM are calculated respectively through the following formulas:
[0060]
[0061] In the formula: j is the corresponding ARIMA prediction algorithm, PROPHET prediction algorithm or LSTM prediction algorithm.
[0062] Based on one of the embodiments, it can be known that the present invention has the following advantages:
[0063] (1) High-precision short-term prediction: By decomposing the total container receiving volume task by route, each route can be analyzed and predicted independently. The characteristics of different routes (such as volatility, trend) can be fully captured, and by combining multiple time series algorithms and a decomposition and integration framework, the present invention can perform high-precision prediction on the container receiving volume in the next 24 hours and flexibly model according to the characteristics of different routes.
[0064] (2) Resource optimization and decision support: The prediction results can help terminal managers plan resource allocation in advance (such as yard vacancies, equipment, manpower, etc.), optimize operation scheduling, and reduce yard congestion and equipment idleness.
[0065] (3) Robustness and stability: Through dynamic weighted integration and outlier elimination, the weights are determined according to the performance of each algorithm on historical data, ensuring that the final prediction results are more accurate. At the same time, the system also screens and eliminates outliers in the prediction results, further improving the stability and reliability of the prediction results and reducing the bias that may be brought by a single algorithm.
[0066] (4) Wide applicability: The present invention not only overcomes the shortcomings of traditional methods in dealing with complex time series, but also significantly improves the accuracy and generalization ability of prediction through a multi-algorithm integration strategy. Through the present invention, terminal managers can allocate resources more reasonably, reduce operation conflicts and resource waste, and improve the overall operation efficiency of the terminal. And the present invention can be applied to container terminals of different scales and various scenarios, providing strong support for the digital transformation and intelligent operation of the terminal.
[0067] The Second Embodiment
[0068] Please refer to Figure 2 , a short-term container receiving volume prediction system for a container terminal gate, including a data acquisition module, a data processing module, an algorithm prediction module, and a result integration module connected in sequence; where
[0069] Data acquisition module: used to obtain the historical container receiving data of the container terminal, including shipping line, container volume, and arrival time. This module automatically extracts data from the terminal data management system and stores it as structured time series data.
[0070] Data processing module: used to preprocess the collected data, including: data cleaning and missing value supplementation, and decomposing the container receiving tasks into time series of container receiving volumes for each shipping line based on the characteristics of the shipping lines.
[0071] Algorithm prediction module: independently predicts the container receiving volumes of each shipping line using the ARIMA algorithm, the PROPHET algorithm, and the LSTM algorithm respectively.
[0072] Result integration module, integrates the prediction results of each shipping line based on the methods of handling outliers, time axis alignment, and weighted average method, and outputs the total container receiving volume in the next 24 hours. By outputting the final prediction result in the form of a time series curve, it provides data support for terminal operation scheduling and resource optimization.
[0073] In this embodiment, in order to better use the method described in the first embodiment, the present application proposes a short-term container receiving volume prediction system for a container terminal gate. Each module corresponds to each step of the above method, and its specific principle has been described above and will not be elaborated here.
[0074] The Third Embodiment
[0075] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the method described in the first embodiment.
[0076] In the above technical solution, in order to better run and use the method described in the first embodiment, the above method is stored in a computer-readable storage medium, and a processor is used to implement the method described in the first embodiment. It should be noted that the principle and effect of each step have been described above and will not be elaborated here.
[0077] Embodiments of the mechanisms disclosed in this application can be implemented in hardware, software, firmware, or a combination of these implementation methods. Embodiments of this application can be implemented as a computer program or program code executed on a programmable system, which includes at least one processor, a storage system (including volatile and non-volatile memories and / or storage elements), at least one input device, and at least one output electronic device.
[0078] The program code can be applied to the input instructions to perform the various functions described in this application and generate output information. The output information can be applied to one or more output electronic devices in a known manner. For the purposes of this application, a processing system includes any system having a processor such as, for example, a digital signal processor (DSP), a microcontroller, an application specific integrated circuit (ASIC), or a microprocessor.
[0079] The program code can be implemented in a high-level procedural language or an object-oriented programming language to communicate with the processing system. When needed, the program code can also be implemented in assembly language or machine language. In fact, the mechanisms described in this application are not limited to the scope of any specific programming language. In any case, the language can be a compiled language or an interpreted language.
[0080] In some cases, the disclosed embodiments can be implemented in hardware, firmware, software, or any combination thereof. The disclosed embodiments can also be implemented as instructions carried or stored on one or more transient or non-transient machine-readable (e.g., computer-readable) storage media, which can be read and executed by one or more processors. For example, the instructions can be distributed via a network or via other computer-readable media. Thus, a machine-readable medium can include any mechanism for storing or transmitting information in a machine (e.g., computer) readable form, including but not limited to, floppy disks, optical disks, optical discs, read-only memories
[0081] (compactdisc-readonlymemory, CD-ROMs), magneto-optical discs, readonlymemory (ROM), randomaccessmemory (RAM), erasableprogrammablereadonlymemory (EPROM), electricallyerasableprogrammableread-onlymemory (EEPROM), magnetic or optical cards, flash memory, or tangible machine-readable memories for transmitting information (e.g., carrier waves, infrared signals, digital signals, etc.) in the form of electrical, optical, acoustic, or other propagated signals using the Internet. Thus, machine-readable media include any type of machine-readable media suitable for storing or transmitting electronic instructions or information in a form readable by a machine (e.g., a computer).
[0082] The above are only some embodiments of the present invention, and thus do not limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall equally be included in the patent protection scope of the present invention.
Claims
1. A method for predicting the short-term container volume received at a container terminal gate, characterized in that: The method comprises: Obtain historical container receiving data at container terminals, including routes, container volumes, and arrival times; Preprocess the container collection data, and decompose the container collection task into a time series of container collection volume of each route based on route characteristics; Use multiple time series prediction algorithms to independently predict the box collection volume of each route, including: ARIMA prediction algorithm, PROPHET prediction algorithm and LSTM prediction algorithm; The prediction results of the above three prediction algorithms for each route are integrated. The integration process includes: Handle outliers and remove extreme prediction values; Align the forecast results of each route on the time axis; The prediction results of the above three prediction algorithms are matched with corresponding weight values by weighted average method; Output the total box collection volume forecast results for the next 24 hours.
2. The method for predicting the short-term container volume received at a container terminal gate according to claim 1 is characterized in that: The preprocessing steps for the inbox data include data cleaning and missing value supplementation.
3. The method for predicting the short-term container volume received at a container terminal gate according to claim 1, characterized in that: The weight values are determined based on the performance of each algorithm on the historical data set by evaluating its mean absolute error or mean square error.
4. The method for predicting the short-term container volume at a container terminal gate according to claim 1, characterized in that: The processing of abnormal values includes setting a threshold region range, and automatically eliminating predicted values that exceed the set threshold region range.
5. The method for predicting the short-term container volume received at a container terminal gate according to claim 1, characterized in that: The output of the total box collection volume forecast within the next 24 hours is calculated using the following formula: Where: n represents the total number of routes, Represents the forecast result calculated by weighted average method Calculated by the following formula: Where: It represents the volume of containers received by route i at time point t predicted by the ARIMA forecasting algorithm. It represents the volume of containers received by route i at time point t predicted by the PROPHET prediction algorithm. It represents the volume of containers received predicted by route i at time point t using the LSTM prediction algorithm; wω A RIMA represents the forecast error of the ARIMA forecasting algorithm; w PROPHET Represents the prediction error of the PROPHET prediction algorithm; w LSTM Represents the prediction error of the LSTM prediction algorithm.
6. The method for predicting the short-term container volume received at a container terminal gate according to claim 5, characterized in that: The w ARIMA 、w PROPHET ,ω LSTM Calculated by the following formulas: Where: j is the corresponding ARIMA prediction algorithm, PROPHET prediction algorithm or LSTM prediction algorithm.
7. A short-term container volume forecasting system for a container terminal gate, characterized in that: The system comprises: The data acquisition module is used to obtain the historical container receiving data of the container terminal, including routes, container volume, and arrival time; A data processing module, used to pre-process the container collection data and decompose the container collection task into a time series of container collection volume of each route based on route characteristics; The algorithm prediction module uses the ARIMA algorithm, PROPHET algorithm and LSTM algorithm to independently predict the box collection volume of each route; The result integration module integrates the forecast results of each route based on processing outliers, time axis alignment and weighted average method, and outputs the total container volume in the next 24 hours.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.