Intelligent piling method and system for non-flow yard

Abstract

The present application relates to the field of logistics management, and discloses a non-flowing yard wisdom pointing and piling method and system, specifically comprising the following steps: obtaining the basic information of the yard, the historical and current cargo arrival and departure information, constructing a cargo prediction model according to the historical cargo information, and obtaining the arrival and departure prediction information of the cargo in a future period of time. A target function based on distance cost, time cost and economic cost is set, the preference degree between the cargo and the yard is generated, and a yard allocation model based on combination optimization is constructed based on this, the yard where the cargo should be stacked is determined, and then an intelligent moving and piling model is constructed, so that the same kind of cargo is stacked in a yard as much as possible through moving and piling in a small area. The present application will help the yard to reduce economic cost and improve operating efficiency.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, and more specifically to a non-process-oriented intelligent stacking method and system for yards. Background Technology

[0002] With the rapid development of the transportation industry, the demand for efficient and precise stacking management is receiving increasing attention, especially in key areas of logistics and transportation such as ports. The layout and operation strategies of the storage yard play a crucial role in the smooth operation of cargo throughout the port.

[0003] Although many yards have been upgraded to flow-oriented yards, which can transport goods directly from berths to stacking locations via conveyor belts and have relatively regular stacking, there are still many non-flow-oriented yards that require dump trucks to transport goods from berths to designated stacking locations. At the same time, due to the complexity of the types of goods and the different stacking rules, their stacking shapes are also more complex. Therefore, compared to flow-oriented yards, the stacking model for non-flow-oriented yards needs to consider more factors and is more challenging.

[0004] Traditional yard allocation often relies on the operation and on-site judgment of technical workers, which may lead to certain limitations, such as incomplete information, low efficiency, and uneven resource allocation. It is gradually failing to meet the management needs of modern ports for yard assignment. With the rapid development of operations research and optimization, it has become possible to develop intelligent allocation models for non-process-oriented yards.

[0005] However, the existing intelligent stacking technology for yards can realize intelligent planning of yards, streamlined stacking of yards, and real-time allocation of stacking positions based on the actual situation on site.

[0006] However, the following drawbacks still exist:

[0007] Firstly, the existing technology mainly relies on manual allocation of yard locations and notification via telephone, which may lead to several problems: (1) When multiple goods are encountered at the same time, it is easy to make real-time allocation impossible; (2) Due to factors such as shift work, communication may be poor, leading to errors in some links; (3) Since real-time allocation by staff is required, it is difficult to make medium- and long-term plans for the yard, so when subsequent goods arrive, they need to be transferred within the yard, resulting in low efficiency.

[0008] Secondly, existing technologies only explain how to achieve intelligent planning of bulk dry cargo terminal yards, but do not specify which constraints are used to achieve intelligent planning, what methods are selected to obtain the results, and the output of the model is only used to assist decision-making and cannot achieve real-time automated allocation of yard locations.

[0009] Thirdly, although existing technologies use the skyline heuristic algorithm and the domain search algorithm to find the optimal solution for mixed loading, this method cannot be well applied to the field of non-process-oriented storage yards. This is because the storage allocation process in non-process-oriented storage yards needs to consider the stacking of subsequent goods, rather than just the current location allocation of these goods. Otherwise, it may lead to the need to transfer the previous goods when the subsequent goods arrive, thus increasing costs.

[0010] Fourth, existing technologies mainly target standardized storage yards and can only be used when the cargo stacking shape is a standard rectangle and the types of cargo are relatively few. They are not suitable for non-standardized storage yards where the stacking shapes are varied and the types of cargo are more complex. Summary of the Invention

[0011] In order to overcome the above-mentioned defects of the prior art, the present invention provides a non-process-oriented intelligent stacking method and system for stacking, so as to solve the problems existing in the background art.

[0012] This invention provides the following technical solution: a non-process-oriented smart stacking method for stockpiling yards, comprising:

[0013] S1: Obtain cargo-related information data, construct a time series prediction model based on ARIMA, and output prediction information for future cargo arrivals and shipments through model training and validation;

[0014] S2: By combining the multi-dimensional matching rules between cargo type and yard with the objective function, the matching degree between each cargo and the yard location is quantified, and a cargo-yard preference matrix is ​​constructed. Each element value in the matrix represents the preference strength of the cargo for the yard location.

[0015] S3: Based on the forecast information of future cargo arrivals and shipments and the cargo-yard preference matrix, construct a yard allocation model based on combinatorial optimization to determine the yards where the cargo should be stored, and output the yard information to be allocated for each batch of cargo from the model.

[0016] S4: Based on the cargo yard allocation information, construct a stacking and repacking model, and output the time and location information of the cargo that needs to be stacked or repacked.

[0017] S5: By combining and analyzing the information on the allocation of goods to the storage yard with the time and location information of the goods that need to be moved or stacked, the storage yard location information of each shipment is obtained and synchronized to each terminal.

[0018] Preferably, the construction and prediction of the ARIMA-based time series forecasting model includes:

[0019] Acquire historical data, current data, and expected arrival data of goods. The data content includes the arrival time, arrival method, dispatch time, and dispatch method of the goods. The historical data is the information data of goods operations completed in the past period. The current data is the real-time or near real-time information data of goods status. The expected arrival data is the information data of goods expected to arrive or be dispatched in the future period.

[0020] Based on the periodicity of cargo transportation, an ARIMA-based time series prediction model is constructed to predict cargo arrivals and departures in the future. By extracting the weight or quantity of cargo arrivals and departures for past and current time periods from historical and current data, arrival and departure time series are formed. The stationarity of the arrival and departure time series is tested. The tested arrival and departure time series, along with relevant external variables, are input as feature variables into the trained ARIMA-based time series prediction model, thereby outputting prediction information for future cargo arrivals and departures. The relevant external variables include collected macroeconomic indicators, industry data, transportation strategy change information, and major event identifiers.

[0021] Historical data is divided into training, validation, and test sets. The constructed ARIMA-based time series prediction model is trained, validated, and tested separately. When goods actually arrive or are shipped, the current data is added to the historical dataset as the latest historical value provided in the rolling forecast and input into the model. The model outputs the predicted value and compares it with the current actual value that occurs subsequently to monitor and evaluate the model performance.

[0022] Predicted data is used as a direct reference for resource planning, and the prediction results are corrected and adjusted after the model prediction.

[0023] Preferably, the construction of the cargo-yard preference matrix includes:

[0024] To set an objective function based on distance cost, time cost, and economic cost, the specific steps include:

[0025] S201: Based on the arrival and dispatch methods of the goods, determine the actual distance from the origin to the destination of the goods, and calculate the distance cost of the arrival and dispatch process by combining the unit distance transportation cost of different transportation methods.

[0026] S202: Calculate time costs based on the dwell time and actual transport time of goods, combined with the value of the goods and the sensitivity to time.

[0027] S203: Calculate the economic cost based on the yard charges, the characteristics of the cargo, and the transport contract;

[0028] S204: Based on actual business conditions and decision-making needs, set different weighting coefficients for distance cost, time cost, and economic cost respectively;

[0029] S205: The objective function is obtained by multiplying the calculated distance cost, time cost, and economic cost by their respective weighting coefficients and then summing them.

[0030] Based on the set objective function, the matching degree between cargo types and storage yards is analyzed, and the preference degree between each cargo and each storage yard is calculated. That is, each element value represents the preference strength of the cargo for the storage yard location, so as to construct the cargo-storage yard preference matrix.

[0031] Preferably, the construction and output of the stockyard allocation model based on combinatorial optimization includes:

[0032] The relevant information of the storage yard and the requirements of the cargo owner company are obtained, and combined with the predicted information of future cargo arrivals and shipments and the cargo-storage yard preference matrix as the features of the model. Together, these features are used to construct the features of the storage yard allocation model based on combinatorial optimization. Then, an integer programming solver is used to solve the model. After the model is solved, the storage yard information to be allocated to each batch of cargo is obtained.

[0033] Preferably, the construction and output of the stacking and repacking model includes:

[0034] Based on the cargo yard allocation information, the objective function is determined and constraints are constructed. The objective function and constraints are combined to form a complete stacking and repacking model. Then, an integer programming solver is used to solve the model. After the model is solved, the time and location information of the cargo that needs to be stacked or repacked are obtained.

[0035] Preferably, the analysis steps for determining the storage location of each shipment include:

[0036] S501: Extract the yard number, yard capacity, current yard usage, and yard geographical location from the information on the yard to which the goods should be allocated. Extract the batch of goods to be moved or reassembled, start time of the move or reassemble, expected end time, original yard location, and target yard location from the time and location information of the goods to be moved or reassembled. Verify, clean, and standardize the format of the extracted data.

[0037] S502: Based on the estimated arrival time of the goods, match the arrival time with the available time of the yard, and select yards that have completed the cleaning of the previous batch of goods and are in a usable state before the arrival time.

[0038] Analyze the sequence and interval between the time of palletizing or stacking operations and the arrival and shipment times of goods, create a timeline chart, and visually detect whether there are time conflicts in the palletizing or stacking operations. If there are time conflicts, adjust according to the priority of goods, prioritize the arrival and shipment time of urgent orders, and appropriately postpone the palletizing operation time of non-urgent goods.

[0039] Based on the layout of the storage yards and transportation nodes, calculate the distance from different storage yards to the loading and unloading area. Then, based on the estimated shipping time of the goods, analyze the relationship between the shipping time and the proximity of the storage yard to the loading and unloading area, and prioritize the storage yards that are closer to the loading and unloading area.

[0040] S503: Calculate the required yard space based on the volume and quantity of the goods, match the calculated required space with the remaining available space of each yard, and select the yard that can accommodate the goods.

[0041] Match cargo types with the characteristics of the selected storage yards, and select the storage yard that matches the cargo type.

[0042] Based on the overall layout of the storage yard and transportation nodes, analyze the cargo flow path planning and select a storage yard that is close to the cargo loading and unloading area and transportation channel;

[0043] S504: Establish an evaluation index system, determine the indicators used to evaluate the merits of yard selection, and then determine the importance of each indicator based on business priorities and actual needs, and set corresponding weights for each indicator.

[0044] Based on the indicators for evaluating the merits of yard selection, a heuristic algorithm or an exact algorithm is selected to find the optimal yard selection scheme. The yard location information for each shipment is directly extracted from the scheme and compiled into a detailed report. Finally, the cargo storage location information is synchronized to each terminal.

[0045] To achieve the above objectives, the present invention provides the following technical solution: a non-process-oriented smart stacking system for stockyards, implementing the aforementioned non-process-oriented smart stacking method for stockyards, comprising:

[0046] Model building and prediction module: Acquire cargo-related information data, build a time series prediction model based on ARIMA, and output prediction information for future cargo arrivals and shipments through model training and validation;

[0047] Cargo-Store Yard Preference Matrix Construction Module: By combining multi-dimensional matching rules between cargo types and storage yards with an objective function, the matching degree between each cargo and the storage yard location is quantified, and a cargo-storey preference matrix is ​​constructed. Each element value in the matrix represents the preference strength of the cargo for the storage yard location.

[0048] Yard allocation analysis module: Based on the forecast information of future cargo arrivals and shipments and the cargo-yard preference matrix, a yard allocation model based on combinatorial optimization is constructed to determine the yards where the cargo should be stored, and the model outputs the yard allocation information for each batch of cargo.

[0049] Stacking or stacking information analysis module: Based on the cargo yard allocation information, a stacking or stacking model is constructed, and the model outputs the time and location information of the cargo that needs to be stacked or stacked.

[0050] Stacking location information output module: By combining and analyzing the information on the storage yard to be allocated to the goods with the time and location information of the goods to be moved or stacked, the storage yard location information of each batch of goods is obtained and synchronized to each terminal.

[0051] The technical effects and advantages of this invention are as follows:

[0052] (1) By combining and analyzing the information on the allocation of goods to the storage yard with the time and location information of the goods to be moved or stacked, the storage yard location information of each shipment is obtained and synchronized to various terminals. The visualization of the goods stacking can be realized through multiple terminals such as mobile APP, so that the cargo owner and relevant operators can clearly and intuitively obtain relevant information such as the storage yard location.

[0053] (2) By combining the multi-dimensional matching rules of cargo type and yard, the matching degree of each cargo and yard location is quantified by the objective function, specific constraints between cargo and yard location are constructed, and a yard allocation model based on combinatorial optimization is constructed to determine the yard where the cargo should be stacked. The model outputs the yard information to be allocated for each batch of cargo, and can select a suitable solution algorithm to obtain the yard allocation result.

[0054] (3) By acquiring relevant cargo information data, a time series prediction model based on ARIMA is constructed. Through model training and verification, prediction information of future cargo arrival and dispatch is output. It is not limited to how to optimally allocate cargo stacking locations under the current circumstances, but also takes into account time factors so that the current allocation results can take into account the optimal allocation of cargo to be delivered later. Through overall planning and consideration, the efficiency can be optimized throughout the entire period.

[0055] (4) Based on the information of the storage yard to be allocated to the goods, construct a stacking and repacking model. The model outputs the time and location information of the goods to be stacked or repacked, and combines it with the information of the storage yard to be allocated to the goods. It can fully consider the stacking methods of various possible storage locations, and is not limited to a certain type of goods or a specific stacking shape. Attached Figure Description

[0056] Figure 1 This is a diagram illustrating the method steps of the present invention.

[0057] Figure 2 This is a system structure block diagram of the present invention. Detailed Implementation

[0058] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The non-process-oriented intelligent stacking method and system of the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] like Figure 1 The embodiment shown provides a non-process-oriented smart stacking method for stockpiling, including:

[0060] S1: Obtain cargo-related information data, construct a time series prediction model based on ARIMA, and output prediction information for future cargo arrivals and shipments through model training and validation.

[0061] In this embodiment, the construction and prediction of the ARIMA-based time series forecasting model includes:

[0062] Acquire historical data, current data, and expected arrival data of goods. The data content includes the arrival time, arrival method, dispatch time, and dispatch method of the goods. The historical data is the information data of goods operations completed in the past period. The current data is the real-time or near real-time information data of goods status. The expected arrival data is the information data of goods expected to arrive or be dispatched in the future period.

[0063] Based on the periodicity of cargo transportation, an ARIMA-based time series prediction model is constructed to predict cargo arrivals and departures in the future. By extracting the weight or quantity of cargo arrivals and departures for past and current time periods from historical and current data, arrival and departure time series are formed. The stationarity of the arrival and departure time series is tested. The tested arrival and departure time series, along with relevant external variables, are input as feature variables into the trained ARIMA-based time series prediction model, thereby outputting prediction information for future cargo arrivals and departures. The relevant external variables include collected macroeconomic indicators, industry data, transportation strategy change information, and major event identifiers.

[0064] Historical data is divided into training, validation, and test sets. The constructed ARIMA-based time series forecasting model is trained, validated, and tested separately. When goods actually arrive or are shipped, the current data is added to the historical dataset as the latest historical value provided in the rolling forecast, and the model outputs a predicted value. The model's predicted value is compared with the subsequent actual value to monitor and evaluate the model's performance. The predicted data is used as a direct reference for resource planning, and the forecast results are corrected and adjusted after the model's prediction.

[0065] It is necessary to explain in detail the data acquisition and prediction:

[0066] S101: Obtain relevant information about the goods: Obtain historical, current and expected arrival data of the goods. The data content includes the arrival time, arrival method, dispatch time, dispatch method, and the arrival and dispatch methods include road, train, ship and other methods.

[0067] S102: Establishing a Prediction Model: Since freight transportation often exhibits periodicity, such as weekly, monthly, and quarterly cycles, this section considers establishing a time series prediction model based on ARIMA. This is used to predict the arrival and dispatch of goods in the future.

[0068] S103: Training the prediction model: The structure of the ARIMA model is determined by the autocorrelation function (ACF), partial autocorrelation function (PACF), and stationarity test (ADF). Then, the specific parameters of the ARIMA model are determined by the minimization information criterion (AIC) and Bayesian information criterion (BIC).

[0069] S104: Test the prediction model: Test the randomness of the residual sequence in the model, i.e., the residuals Whether the residual sequence exhibits randomness can be tested using the autocorrelation function method. If it exhibits randomness, the model is reasonable; if it lacks randomness, the model is unreasonable.

[0070] S105: Output the prediction results of the prediction model: Based on the ARIMA-based time series prediction model obtained through training, the arrival time, arrival method, dispatch time, and dispatch method of the goods can be obtained, which will help to make the subsequent yard allocation more reasonable.

[0071] S2: By combining the multi-dimensional matching rules between cargo type and yard with the objective function, the matching degree between each cargo and the yard location is quantified, and a cargo-yard preference matrix is ​​constructed. Each element value in the matrix represents the preference strength of the cargo for the yard location.

[0072] In this embodiment, the construction of the cargo-yard preference matrix includes:

[0073] To set an objective function based on distance cost, time cost, and economic cost, the specific steps include:

[0074] S201: Based on the arrival and dispatch methods of the goods, determine the actual distance from the origin to the destination of the goods, and calculate the distance cost of the arrival and dispatch process by combining the unit distance transportation cost of different transportation methods.

[0075] S202: Calculate time costs based on the dwell time and actual transport time of goods, combined with the value of the goods and the sensitivity to time.

[0076] S203: Calculate the economic cost based on the yard charges, the characteristics of the cargo, and the transport contract;

[0077] S204: Based on actual business conditions and decision-making needs, set different weighting coefficients for distance cost, time cost, and economic cost respectively;

[0078] S205: The objective function is obtained by multiplying the calculated distance cost, time cost, and economic cost by their respective weighting coefficients and then summing them.

[0079] Based on the set objective function, the matching degree between cargo types and storage yards is analyzed, and the preference degree between each cargo and each storage yard is calculated. That is, each element value represents the preference strength of the cargo for the storage yard location, so as to construct the cargo-storage yard preference matrix.

[0080] It should be specifically noted that, based on the cost factors involved in the objective function, the matching between cargo types and storage yards is analyzed in depth from the dimensions of cargo characteristics, transportation mode, and cost-effectiveness. Based on this, the preference degree between each cargo and each storage yard is calculated. ,in Represents positive numbers. Represents the objective function value. Indicates goods, Let M represent the location of the storage yard, and let N represent the preference strength of the cargo for that location. Assuming there are M cargoes and N storage yard locations, the preference values ​​calculated for each cargo and each storage yard location are used as matrix elements to construct a matrix. The cargo-yard preference matrix is ​​in the following form:

[0081]

[0082] in, Indicates goods For the location of the storage yard The matrix shows the preference strength for each cargo to different yard locations.

[0083] S3: Based on the forecast information of future cargo arrivals and shipments and the cargo-yard preference matrix, construct a yard allocation model based on combinatorial optimization to determine the yards where the cargo should be stored, and output the yard information to be allocated to each batch of cargo from the model.

[0084] In this embodiment, the construction and output of the stockyard allocation model based on combinatorial optimization includes:

[0085] The relevant information of the storage yard and the requirements of the cargo owner company are obtained, and combined with the predicted information of future cargo arrivals and shipments and the cargo-storage yard preference matrix as the features of the model. Together, these features are used to construct the features of the storage yard allocation model based on combinatorial optimization. Then, an integer programming solver is used to solve the model. After the model is solved, the storage yard information to be allocated to each batch of cargo is obtained.

[0086] It should be specifically noted that, based on the previously determined cargo and yard preference matrix, and taking into account actual factors such as the cargo owner's requirements and the specific attributes of the cargo, and using the maximum area that each yard can accommodate as a constraint, the model constraints are designed to reflect the actual situation on site to the greatest extent possible. Therefore, the specific steps for constructing the yard allocation model based on the above process include:

[0087] S301: Variable Description:

[0088] S302: Constructing Model Constraints:

[0089] (1) For any storage yard, ensure that the area used for goods does not exceed the upper limit of the available area:

[0090]

[0091] (2) For any given item, ensure that all items can be stacked:

[0092]

[0093] (3) For any cargo in any yard, ensure that the area and weight of the cargo correspond:

[0094]

[0095] The correspondence between i and k is determined based on the type of goods.

[0096] (4) The stacking situation conforms to the actual situation:

[0097] Under normal circumstances, a batch of goods is concentrated in one storage yard, and at most two storage yards.

[0098]

[0099] (5) Preference constraint: the preference of the stackable storage yard is not 0.

[0100]

[0101] (6) The start and end times of the storage yard meet the requirements for berthing and departure:

[0102]

[0103]

[0104]

[0105] S303: Define the objective function:

[0106] (1) Distance factor:

[0107] This includes the distance from the berth to the yard and the distance from the yard to the shipping point;

[0108] (2) Preference factors:

[0109] By assigning different weights, goods can be placed in suitable areas;

[0110] (3) Factors related to stacking:

[0111] Due to practical limitations, we aim to avoid stacking as much as possible. Therefore, we add a penalty to minimize the occurrence of stacking separation.

[0112] S304: Incorporation of Time Element:

[0113] The above constraints hold true from the current moment to any future moment, and the model is iteratively solved using a mathematical solver to obtain the optimal result;

[0114] S305: Verify model performance:

[0115] Use historical data to validate and evaluate the model's performance, calculate the difference between the model's output and the actual historical data, analyze the possible reasons behind the difference, and iteratively optimize the model accordingly.

[0116] S306: Output the results of the stockyard allocation model:

[0117] Based on the yard allocation model of combinatorial optimization, the yard allocation information of any cargo at any given time can be obtained in real time.

[0118] S4: Based on the cargo yard allocation information, construct a stacking and repacking model, and output the time and location information of the cargo that needs to be stacked or repacked.

[0119] In this embodiment, the construction and output of the stacking and repacking model includes:

[0120] Based on the cargo yard allocation information, the objective function is determined and constraints are constructed. The objective function and constraints are combined to form a complete stacking and repacking model. Then, an integer programming solver is used to solve the model. After the model is solved, the time and location information of the cargo that needs to be stacked or repacked are obtained.

[0121] It should be specifically noted that, due to practical limitations, goods should be stored in the same storage yard as much as possible, with a maximum of two storage yards. Therefore, when a storage yard cannot meet the stacking requirements, it is necessary to identify some smaller stacks and consider shifting or merging them so that the original area can be used to store other goods intact. The specific steps for constructing a shifting and merging model based on the above process include:

[0122] S401: Obtain existing data and initialize stacking data:

[0123] Obtain the available area of ​​each storage yard and the projected future storage area required. Based on historical data, determine whether currently stored goods meet the criteria for consolidation, i.e., the same owner and type of goods, and designate these goods as consolidable.

[0124] S402: Define the objective function:

[0125] (1) Distance factor:

[0126] The distance between stacks being moved or combined;

[0127] (2) Preference factors:

[0128] When selecting a new storage area for re-stacking, the preference level between the goods and the new storage area needs to be considered. This factor is not required for consolidation, as it necessarily meets the preference requirement.

[0129] (3) Value cost factors:

[0130] Calculate the value cost of cargo transportation and minimize stacking and repacking to save costs;

[0131] (4) Factors affecting area variation:

[0132] By shifting or combining stacks, the stackable area needs to be increased to meet the subsequent stacking area requirements for goods.

[0133] S403: Constructing Model Constraints:

[0134] For each storage yard, iterate through areas with an area smaller than a fixed value to determine whether the stacking and relocation plan can be executed. Define decision variables (0-1 variables). When the variable is 0, it does not move; when it is 1, it moves from... The storage yard was moved to Storage yard. Simultaneously define decision variables (floating-point variables) representing the storage yard situation after stacking and repacking. After stacking and repacking, the area of ​​the goods in the corresponding storage yard needs to be reduced or increased, and it must be ensured that the goods are only moved to one stacking location.

[0135]

[0136]

[0137]

[0138]

[0139]

[0140] S404: Verify model performance:

[0141] The rationality of the stack transfer and re-stacking scheme is determined by evaluating the output of the model's objective function, analyzing potential problems, and iteratively optimizing the model accordingly.

[0142] S405: Output the results of the stack transfer and stacking model:

[0143] Based on the actual constraints of the stacking and repacking model, the time and location information of goods that need to be stacked or repacked can be obtained.

[0144] S5: By combining and analyzing the information on the allocation of goods to the storage yard with the time and location information of the goods that need to be moved or stacked, the storage yard location information of each shipment is obtained and synchronized to each terminal.

[0145] In this embodiment, the analysis steps for determining the storage location of each shipment include:

[0146] S501: Extract the yard number, yard capacity, current yard usage, and yard geographical location from the information on the yard to which the goods should be allocated. Extract the batch of goods to be moved or reassembled, start time of the move or reassemble, expected end time, original yard location, and target yard location from the time and location information of the goods to be moved or reassembled. Verify, clean, and standardize the format of the extracted data.

[0147] S502: Based on the estimated arrival time of the goods, match the arrival time with the available time of the yard, and select yards that have completed the cleaning of the previous batch of goods and are in a usable state before the arrival time.

[0148] Analyze the sequence and interval between the time of palletizing or stacking operations and the arrival and shipment times of goods, create a timeline chart, and visually detect whether there are time conflicts in the palletizing or stacking operations. If there are time conflicts, adjust according to the priority of goods, prioritize the arrival and shipment time of urgent orders, and appropriately postpone the palletizing operation time of non-urgent goods.

[0149] Based on the layout of the storage yards and transportation nodes, calculate the distance from different storage yards to the loading and unloading area. Then, based on the estimated shipping time of the goods, analyze the relationship between the shipping time and the proximity of the storage yard to the loading and unloading area, and prioritize the storage yards that are closer to the loading and unloading area.

[0150] S503: Calculate the required yard space based on the volume and quantity of the goods, match the calculated required space with the remaining available space of each yard, and select the yard that can accommodate the goods.

[0151] Match cargo types with the characteristics of the selected storage yards, and select the storage yard that matches the cargo type.

[0152] Based on the overall layout of the storage yard and transportation nodes, analyze the cargo flow path planning and select a storage yard that is close to the cargo loading and unloading area and transportation channel;

[0153] S504: Establish an evaluation index system, determine the indicators used to evaluate the merits of yard selection, and then determine the importance of each indicator based on business priorities and actual needs, and set corresponding weights for each indicator.

[0154] Based on the indicators for evaluating the merits of yard selection, a heuristic algorithm or an exact algorithm is selected to find the optimal yard selection scheme. The yard location information for each shipment is directly extracted from the scheme and compiled into a detailed report. Finally, the cargo storage location information is synchronized to each terminal.

[0155] It should be noted that the specific steps of the output display and feedback module include:

[0156] The model described above is used to calculate the optimal storage location for each shipment and synchronize this information to all terminals. The central control screen in the control center displays the real-time status of each storage yard, providing detailed information on the status of each shipment and allowing users to switch between past and future storage conditions. The smart terminal app used by storage yard operators provides access to the current and future status of each storage yard, as well as details of tasks to be performed, such as arrival method, arrival time, shipping method, and shipping time. This allows on-site operators to stay informed about storage yard information and assigned tasks. Similarly, the shipper's mobile app provides detailed information on their shipment, including arrival method, arrival time, shipping method, and shipping time, enabling them to understand the status of their goods and make necessary arrangements. Furthermore, the control center should have the authority to modify feedback, allowing modifications to various constraints and output results to meet specific needs.

[0157] like Figure 2 The embodiment shown provides an implementation system corresponding to a non-process-oriented smart stacking method for yards, including a model building and prediction module, a cargo-yard preference matrix construction module, a yard allocation analysis module, a stacking or stacking information analysis module, and a storage location information output module. The model building and prediction module is connected to the yard allocation analysis module, the cargo-yard preference matrix construction module is connected to the yard allocation analysis module, the yard allocation analysis module is connected to the stacking or stacking information analysis module, the stacking or stacking information analysis module is connected to the storage location information output module, and the yard allocation analysis module is connected to the storage location information output module.

[0158] The model building and prediction module acquires cargo-related information data, builds a time series prediction model based on ARIMA, and outputs prediction information for future cargo arrivals and shipments through model training and validation.

[0159] The cargo-yard preference matrix construction module constructs a cargo-yard preference matrix by combining multi-dimensional matching rules between cargo types and yards with an objective function to quantify the matching degree between each cargo and the yard location. Each element value in the matrix represents the preference intensity of the cargo for the yard location.

[0160] The yard allocation analysis module constructs a yard allocation model based on combinatorial optimization based on the predicted information of future cargo arrivals and shipments and the cargo-yard preference matrix, determines the yard where the cargo should be stored, and outputs the yard information to be allocated to each batch of cargo from the model.

[0161] The stacking or stacking information analysis module constructs a stacking or stacking model based on the cargo yard allocation information, and outputs the time and location information of the cargo that needs to be stacked or stacked.

[0162] The stacking location information output module analyzes the information on the storage yard to be allocated to the goods and the time and location information of the goods to be moved or stacked, and obtains the storage yard location information for each shipment, and synchronizes it to each terminal.

[0163] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0164] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A non-process-oriented intelligent stacking method for stockpiling yards, characterized in that, include: S1: Obtain cargo-related information data, construct a time series prediction model based on ARIMA, and output prediction information for future cargo arrivals and shipments through model training and validation; S2: By combining the multi-dimensional matching rules between cargo type and yard with the objective function, the matching degree between each cargo and the yard location is quantified, and a cargo-yard preference matrix is ​​constructed. Each element value in the matrix represents the preference strength of the cargo for the yard location. S3: Based on the forecast information of future cargo arrivals and shipments and the cargo-yard preference matrix, construct a yard allocation model based on combinatorial optimization to determine the yards where the cargo should be stored, and output the yard information to be allocated for each batch of cargo from the model. S4: Based on the cargo yard allocation information, construct a stacking and repacking model, and output the time and location information of the cargo that needs to be stacked or repacked. S5: By combining and analyzing the information on the allocation of goods to the storage yard with the time and location information of the goods that need to be moved or stacked, the storage yard location information of each shipment is obtained and synchronized to each terminal. The construction of the cargo-yard preference matrix includes: To set an objective function based on distance cost, time cost, and economic cost, the specific steps include: S201: Based on the arrival and dispatch methods of the goods, determine the actual distance from the origin to the destination of the goods, and calculate the distance cost of the arrival and dispatch process by combining the unit distance transportation cost of different transportation methods. S202: Calculate time costs based on the dwell time and actual transport time of goods, combined with the value of the goods and the sensitivity to time. S203: Calculate the economic cost based on the yard charges, the characteristics of the cargo, and the transport contract; S204: Based on actual business conditions and decision-making needs, set different weighting coefficients for distance cost, time cost, and economic cost respectively; S205: The objective function is obtained by multiplying the calculated distance cost, time cost, and economic cost by their respective weighting coefficients and then summing them. Based on the set objective function, the matching degree between cargo types and storage yards is analyzed, and the preference degree between each cargo and each storage yard is calculated. That is, each element value represents the preference strength of the cargo for the storage yard location, so as to construct the cargo-storage yard preference matrix. The construction and output of the stockyard allocation model based on combinatorial optimization include: The relevant information of the storage yard and the requirements of the cargo owner company are obtained, and combined with the predicted information of future cargo arrival and shipment and the cargo-storage yard preference matrix as the features of the model. Together, the features of the storage yard allocation model based on combinatorial optimization are constructed. Then, an integer programming solver is used to solve the model. After the model is solved, the storage yard information to be allocated to each batch of cargo is obtained. The construction and output of the stacking and repacking model include: Based on the cargo yard allocation information, an objective function is determined, which includes distance factors, preference factors, value cost factors, and area change factors. Constraints are then constructed, including traversal judgment, decision variable setting, yard area adjustment, and movement restrictions. The objective function and constraints are combined to form a complete stacking and repacking model. An integer programming solver is then used to solve the model. After the model is solved, the time and location information of the cargo that needs to be stacked or repacked are obtained.

2. The non-process-oriented intelligent stacking method for stockpiling yards according to claim 1, characterized in that, The construction and prediction of the ARIMA-based time series forecasting model include: The system acquires historical data, current data, and expected arrival data for goods. The data includes the arrival time, arrival method, dispatch time, and dispatch method of the goods. The historical data is information on goods operations completed within a past period. The current data is real-time or near-real-time information on the status of the goods. The expected arrival data is information on confirmed and deterministic goods arrival or dispatch plans within a future period. Based on the periodicity of freight transportation, an ARIMA-based time series forecasting model is constructed to predict the arrival and departure of goods in the future. By extracting the weight or quantity of goods arriving and departing in past and current time periods from historical and current data, arrival and departure time series are formed. The stationarity of the arrival and departure time series is tested. The tested arrival and departure time series, along with relevant external variables, are input as feature variables into the trained ARIMA-based time series forecasting model, thereby outputting the forecast information of future freight arrival and departure. The relevant external variables include collected macroeconomic indicators, industry data, information on changes in transportation policies, and indicators of major events. Historical data is divided into training, validation, and test sets. The constructed ARIMA-based time series prediction model is trained, validated, and tested separately. When goods actually arrive or are shipped, the current data is added to the historical dataset as the latest historical value provided in the rolling forecast and input into the model. The model outputs the predicted value and compares it with the current actual value that occurs subsequently to monitor and evaluate the model performance. Predicted data is used as a direct reference for resource planning, and the prediction results are corrected and adjusted after the model prediction.

3. The non-process-oriented intelligent stacking method for stockpiling yards according to claim 1, characterized in that, The analysis steps for determining the storage location for each shipment include: S501: Extract the yard number, yard capacity, current yard usage, and yard geographical location from the information on the yard to which the goods should be allocated. Extract the batch of goods to be moved or reassembled, start time of the move or reassemble, expected end time, original yard location, and target yard location from the time and location information of the goods to be moved or reassembled. Verify, clean, and standardize the format of the extracted data. S502: Based on the estimated arrival time of the goods, match the arrival time with the available time of the yard, and select yards that have completed the clearing of the previous batch of goods and are in an available state before the arrival time. Analyze the sequence and interval between the time of palletizing or stacking operations and the arrival and shipment times of goods, create a timeline chart, and visually detect whether there are time conflicts in the palletizing or stacking operations. If there are time conflicts, adjust according to the priority of goods, prioritize the arrival and shipment time of urgent orders, and appropriately postpone the palletizing operation time of non-urgent goods. Based on the location of the storage yard in the overall layout of the port area, calculate the distance from different storage yards to various loading / unloading / transport nodes. Then, based on the expected shipment time of the goods, analyze the relationship between the shipment time and the proximity of the storage yard to the loading / unloading area, and give priority to storage yards that are closer to the loading / unloading area. S503: Calculate the required yard space based on the volume and quantity of the goods, match the calculated required space with the remaining available space of each yard, and select the yard that can accommodate the goods. Match cargo types with the characteristics of the selected storage yards, and select the storage yard that matches the cargo type. S504: Based on the analysis results of S501 to S503, establish an evaluation index system and determine the indicators used to evaluate the merits of yard selection. The evaluation index system includes the following indicators: analysis results of the sequence and interval between the time of pallet transfer or pallet merging operations and the arrival and dispatch time of goods; analysis results of the intuitive detection of whether there are time conflicts in pallet transfer or pallet merging operations; analysis results of the calculation of the distance from different yards to various loading / unloading / transportation nodes; analysis results of matching the space required by the goods with the remaining available space of each yard; analysis results of matching the type of goods with the characteristics of the selected yards; and analysis results of the analysis of the planning of the flow path of goods and the selection of yards that are closer to the loading / unloading area and transportation channel. Then, the importance of each indicator is determined according to the business focus and actual needs, and a corresponding weight is set for each indicator. Based on the indicators for evaluating the merits of yard selection, a heuristic algorithm or an exact algorithm is selected to find the optimal yard selection scheme. The yard location information for each shipment is directly extracted from the scheme and compiled into a detailed report. Finally, the cargo storage location information is synchronized to each terminal.

4. A non-process-oriented intelligent stacking system for a stockyard, implementing the non-process-oriented intelligent stacking method for a stockyard as described in any one of claims 1-3, characterized in that, include: Model building and prediction module: Acquire cargo-related information data, build a time series prediction model based on ARIMA, and output prediction information for future cargo arrivals and shipments through model training and validation; Cargo-Store Yard Preference Matrix Construction Module: By combining multi-dimensional matching rules between cargo types and storage yards with an objective function, the matching degree between each cargo and the storage yard location is quantified, and a cargo-storey preference matrix is ​​constructed. Each element value in the matrix represents the preference strength of the cargo for the storage yard location. Yard allocation analysis module: Based on the forecast information of future cargo arrivals and shipments and the cargo-yard preference matrix, a yard allocation model based on combinatorial optimization is constructed to determine the yards where the cargo should be stored, and the model outputs the yard allocation information for each batch of cargo. Stacking or stacking information analysis module: Based on the cargo yard allocation information, a stacking or stacking model is constructed, and the model outputs the time and location information of the cargo that needs to be stacked or stacked. Stacking location information output module: By combining and analyzing the information on the storage yard to be allocated to the goods with the time and location information of the goods to be moved or stacked, the storage yard location information of each batch of goods is obtained and synchronized to each terminal.

Images