A sea shipping management method for sea shipping assembled containers based on an internet of things platform

Abstract

The application discloses a kind of sea shipping assembly box shipping management methods based on internet of things platform, belong to logistics management technical field.Method includes: obtaining cargo list and extracting cargo attribute, analyzing chemical attribute and physical characteristic, generate conflict risk assessment report;Based on multistage time data construction time limit prediction model, optimize loading priority;Optimize loading scheme by space allocation and dynamic adjustment algorithm, ensure the compatibility of goods and transport timeliness;Real-time monitoring shipping state and according to delay risk trigger emergency adjustment, ensure the execution of transport plan.The present application can effectively improve loading efficiency, guarantee transport safety, reduce transport risk and delay probability, comprehensively improve logistics management level.The present application is suitable for intelligent logistics, sea shipping assembly box transportation and other fields, and provides efficient and accurate decision support for sea shipping management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics management, in particular to a sea shipping container assembly management method based on an Internet of Things platform. This method uses Internet of Things technology, data analysis, optimization algorithms, and real-time monitoring to improve the loading efficiency, safety, and timeliness of sea shipping container transportation. It is suitable for sea shipping, logistics management, and intelligent transportation systems and related fields. BACKGROUND

[0002] With the growth of global trade and logistics demand, sea shipping containers, as one of the most important means of cargo transportation, play a crucial role. However, existing sea shipping container shipping management methods still face challenges in many aspects. Traditional shipping management systems often rely on manual experience or single data processing methods, making it difficult to cope with complex changes and dynamic demands in the shipping process, resulting in low loading efficiency, high safety risks, and inaccurate timeliness prediction.

[0003] Currently, the development of sea shipping container shipping management systems has gone through a transition from manual operation to information management. With the rise of new technologies such as the Internet of Things, artificial intelligence, and big data, intelligent management has begun to be applied in the shipping field. For example, data collection and real-time monitoring based on the Internet of Things technology can effectively improve the accuracy of data in the shipping process. However, existing methods still lack comprehensive solutions in data processing, risk assessment, and loading optimization, especially in how to handle multi-source time data, how to accurately assess cargo risks, and how to optimize loading plans, which still have a large technical gap.

[0004] Therefore, the purpose of the present application is to solve the problems of non-optimal loading, inaccurate risk assessment, and inaccurate timeliness prediction in existing sea shipping container shipping management methods, and to provide an intelligent sea shipping container shipping management method based on an Internet of Things platform to improve transportation efficiency, reduce delays, and ensure transportation safety. SUMMARY

[0005] The present application provides a sea shipping container shipping management method based on an Internet of Things platform, which aims to optimize the loading and transportation process of sea shipping containers by integrating multi-source data, intelligent algorithms, and real-time monitoring, thereby improving transportation efficiency, safety, and timeliness.

[0006] The present application provides a sea shipping container shipping management method based on an Internet of Things platform, comprising:

[0007] Step 1, obtain the cargo list, extract the chemical properties and physical characteristics data of each batch of cargo, form an initial attribute data set, use a classification database to classify the cargo, and obtain the attribute feature set of the cargo;

[0008] Step 2: Analyze the interaction between chemical properties and physical characteristics based on the attribute feature set. If a chemical reaction or physical damage is detected, generate a conflict risk assessment report.

[0009] Step 3: Extract the high-risk cargo combination list from the conflict risk assessment report. Use a dynamic constraint verification mechanism to screen the cargo matching scheme and obtain a preliminary loading scheme.

[0010] Step 4: Obtain the time data of ship navigation, port loading and unloading, and customs clearance. Construct a real-time time data set and generate a comprehensive time limit prediction model based on it.

[0011] Step 5: According to the comprehensive time limit prediction model, sort the delivery time limit requirements of each batch of goods. Use an automatic weight adjustment algorithm to optimize the loading priority and obtain a loading priority list.

[0012] Step 6: Extract the highest priority goods from the loading priority list. Combine the preliminary loading scheme for dynamic adjustment. If a priority conflict is detected, reassign the container position to obtain the final loading scheme.

[0013] Step 7: Real-time monitoring of ship navigation and port operation status. If a delay risk is detected, trigger the emergency process, adjust the loading priority or reassign the container, and obtain the adjusted loading scheme.

[0014] Step 8: Extract the adjustment record from the adjusted loading scheme. Verify it with the real-time time data set to determine whether it meets the time limit requirements of all goods and obtain the verification result.

[0015] Step 9: Generate a container loading plan and scheduling report based on the verification result.

[0016] The technical scheme provided by the embodiment of the application can include the following beneficial effects:

[0017] 1. Through accurate cargo attribute analysis and risk assessment, the loading scheme of the goods is optimized, ensuring that high-risk goods are reasonably handled and avoiding safety accidents caused by improper loading, thereby improving the loading efficiency and safety.

[0018] 2. By integrating time data from multiple links and predicting time limits based on big data and machine learning algorithms, the transportation time limit of each batch of goods can be more accurately mastered, thereby reasonably arranging the operation of ship navigation, port loading and unloading, customs clearance and other links, and improving the overall transportation time efficiency.

[0019] 3. Real-time monitoring of risk factors during transportation, such as ship delays or port operation anomalies, and dynamic optimization through emergency adjustment processes ensure that timeliness and safety are not affected by unforeseen circumstances during transportation, thereby enhancing the flexibility and responsiveness of logistics management.

[0020] 4. Based on the comprehensive time-limit prediction model and automatic weight adjustment algorithm, this invention can optimize the container loading sequence and priority, effectively avoid resource waste, reduce delays caused by unreasonable loading, and improve the utilization rate of logistics resources.

[0021] 5. It can acquire data in real time, predict time limits, optimize loading plans and make emergency adjustments, realizing comprehensive intelligent management in the shipping process of LCL (Less than Container Load) shipping, which greatly improves the efficiency and accuracy of the entire transportation process. Attached Figure Description

[0022] Figure 1 This is a flowchart of a shipping management method for consolidation containers based on an Internet of Things (IoT) platform according to the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0024] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of a shipping management method for consolidation containers based on an Internet of Things (IoT) platform in this application includes:

[0025] Step 1: Obtain the goods list, extract the chemical and physical properties data of each batch of goods to form an initial attribute dataset, and use a classification database to classify the goods to obtain the attribute feature set of the goods.

[0026] In one specific embodiment, the process of performing step 1 may specifically include the following steps:

[0027] Based on the classification database, the initial attribute dataset is matched, and the K-nearest neighbor algorithm is used to determine the preliminary classification of each batch of goods, resulting in a classification result set.

[0028] For the classification result set, principal component analysis algorithm is used to extract the main features of chemical properties and physical characteristics to form a simplified feature set;

[0029] Determine whether the matching degree between the simplified feature set and the preset rules in the classification database is higher than the first preset value. If so, determine the attribute feature set through attribute analysis. If not, re-extract data from the initial attribute dataset to obtain the updated attribute dataset.

[0030] Based on the updated attribute dataset, the support vector machine algorithm is used to perform secondary classification of the goods batches to obtain an optimized classification result set.

[0031] Based on the correlation analysis between the optimized classification result set and the simplified feature set, the attribute feature set of each batch of goods is determined, and the final feature description set is obtained.

[0032] Based on the final feature description set, generate classification labels and attribute feature maps for each batch of goods, and obtain the attribute feature set.

[0033] Specifically, after obtaining the cargo list, the chemical properties and physical characteristics of each batch of goods are extracted by parsing the list structure, including chemical composition, density, volume, weight, etc.

[0034] In the data matching stage, based on a pre-set classification database, the K-nearest neighbor algorithm is used to calculate the distance between each batch of goods and existing categories in the database to determine the similarity between the goods and each category, thereby assigning them to the corresponding category for preliminary classification. To optimize the classification results and extract key features, principal component analysis (PCA) is used to reduce the dimensionality of the chemical properties and physical characteristics in the dataset, extracting the most representative principal components to obtain a simplified feature set. This simplified feature set contains the main features that can maximize the explanation of data variance.

[0035] Based on a simplified feature set, the matching degree with preset rules in the classification database is calculated. If the matching degree is higher than a first preset value, the attribute feature set of the goods is determined through attribute analysis. If the matching degree is lower than the preset value, it indicates that the preliminary classification result may have errors or insufficient features, and the updated attribute dataset needs to be extracted again. For example, the preset rules in the classification database are: Rule 1: If the chemical activity is greater than 0.8 and the temperature stability is higher than 70℃, it is a "high-risk chemical"; Rule 2: If the chemical activity is between 0.5 and 0.8 and the temperature stability is higher than 60℃, it is a "medium-risk chemical"; Rule 3: If the chemical activity is lower than 0.5 and the temperature stability is greater than or equal to 60℃, it is a "low-risk chemical"; Rule 4: If the temperature stability is higher than 100℃, it is a "super-high temperature substance". Taking cargo A01 as an example, its simplified feature set is chemical activity 0.85 and temperature stability 75℃. The simplified feature set of A01 matches Rule 1 100%, meeting the requirements of Rule 1, and belongs to "high-risk chemicals". For other cargoes, such as B02, its simplified feature set is chemical activity 0.60 and temperature stability 90℃, which meets the requirements of Rule 2, but the matching degree with Rule 2 does not reach the first preset threshold. Therefore, attribute data needs to be extracted again and secondary classification needs to be performed.

[0036] The updated attribute dataset is classified based on the new data, and a Support Vector Machine (SVM) algorithm is used for secondary classification of the goods. The SVM algorithm constructs a hyperplane to classify data points into different categories and effectively handles complex nonlinear relationships, thereby optimizing the classification result set. Through correlation analysis with the simplified feature set, the attribute feature set for each batch of goods is finally determined, and a complete feature description set is generated. The feature description set contains the specific classification label and attribute feature mapping for each batch of goods, ensuring that the attribute information of the goods is accurately recorded and managed.

[0037] By combining the K-nearest neighbors algorithm with principal component analysis and secondary classification using support vector machines, the core features of cargo can be effectively extracted, improving classification accuracy and ensuring the accuracy of subsequent steps such as risk assessment and loading optimization. Furthermore, by simplifying the feature set and implementing multi-level classification optimization, redundant information is reduced and computational efficiency is improved. In practical applications, this significantly enhances the intelligence level of the entire LCL (Less than Container Load) shipping management process, ensuring the safety and timeliness of transportation.

[0038] Step 2: Analyze the interaction between chemical properties and physical characteristics based on the attribute feature set. If a potential chemical reaction or physical damage is detected, a conflict risk assessment report is generated.

[0039] In one specific embodiment, the process of performing step 2 may specifically include the following steps:

[0040] Obtain the chemical composition and physical state data of the cargo from the attribute feature set;

[0041] If the reactivity of the chemical components exceeds the second preset value, the chemical reaction path is calculated using a thermodynamic model, and the reaction risk level is determined.

[0042] Based on the reaction risk level, the corresponding temperature threshold and pressure conditions are determined, and finite element analysis is used to simulate the changes in physical properties to obtain stress distribution results.

[0043] If the stress distribution results show that the mechanical strength is lower than the third preset value, then the corrosion tendency and thermal stability are analyzed through molecular dynamics simulation to determine the risk of physical damage.

[0044] Based on the analysis results of corrosion tendency and thermal stability, key features are extracted, and support vector machine is used to classify the risk level to obtain the classified risk assessment results.

[0045] Based on the risk assessment results, density distribution and surface tension data are obtained, and the comprehensive impact intensity of each risk factor is calculated using the numerical integration method, thereby determining the conflict risk level.

[0046] Based on the conflict risk level, a conflict risk assessment report is generated that includes the risks of chemical reactions and physical damage.

[0047] Specifically, chemical composition and physical state data of goods are extracted from the attribute feature set. The chemical components involved include sodium chloride, sodium nitrate, etc., and physical properties such as density, volume, and weight.

[0048] Thermodynamic models help determine the potential risks of reactions by calculating the probability and rate of reaction of substances under different environmental conditions. If the chemical components of the cargo have high reactivity, exceeding a second preset value, the thermodynamic model calculates the chemical reaction pathway and assesses the risk level of the reaction. If the reaction risk level exceeds the preset safety range, the potential threats of the substance during transportation are further analyzed.

[0049] Finite element analysis (FEM) is used to simulate the changes in the physical properties of materials under different temperatures and pressures, particularly stress distribution. FEM establishes a mathematical model of the material under specific conditions and calculates its stress state under external pressure. If the stress distribution results show that the mechanical strength is lower than a third preset value, it indicates insufficient mechanical strength, and the material may suffer physical damage during transportation. Molecular dynamics simulation is then used to analyze the material's corrosion tendency and thermal stability. Molecular dynamics simulation predicts the stability of the material under different environmental conditions by simulating molecular-level interactions, and combines corrosion tendency and thermal stability to determine the magnitude of the physical damage risk. Based on the analysis results of corrosion tendency and thermal stability, key features are extracted, and support vector machine (SVM) is used to classify the risk level. The SVM algorithm finds the optimal separating hyperplane to divide the data into different risk levels, providing accurate classification results.

[0050] Physical property data such as density and surface tension, which are related to risk, are extracted from the set of property characteristics of the goods. This data can be obtained through experiments, material databases, or other means, specifically including the density value (unit: kg / m³) and surface tension (unit: N / m) of each batch of goods. After obtaining this physical property data, the chemical reaction risk and physical damage risk level of each batch of goods are combined, and a weighted calculation of all risk factors is performed using a numerical integration method to obtain the comprehensive impact intensity of each batch of goods. Specifically, the comprehensive impact intensity is calculated using the following formula: Comprehensive Impact Intensity = w1 × Chemical Reaction Risk + w2 × Physical Damage Risk + w3 × Density + w4 × Surface Tension, where w1, w2, w3, and w4 are the weights of each risk factor, representing their contribution to the overall conflict risk. The weights are determined based on their actual impact on the conflict risk. Through this formula, the four factors of chemical reaction risk, physical damage risk, density, and surface tension are comprehensively calculated to obtain the comprehensive impact intensity of each batch of goods.

[0051] For example, suppose the risk assessment results for cargo A01 are: high risk of chemical reaction, low risk of physical damage, density of 2200 kg / m³, and surface tension of 72.5 N / m. Assuming a weighting of 0.4 for chemical reaction risk, 0.3 for physical damage risk, 0.2 for density, and 0.1 for surface tension, the overall impact intensity of A01 can be calculated as 447.63 using the above formula.

[0052] Conflict risk levels can be classified into different grades based on the magnitude of the overall impact intensity. For example: overall impact intensity > 500, high conflict risk; 300 < overall impact intensity ≤ 500, medium conflict risk; overall impact intensity ≤ 300, low conflict risk. According to this standard, the overall impact intensity of cargo A01 is 447.63, which falls under the medium conflict risk level. The conflict risk level of other cargoes can be determined similarly.

[0053] The aforementioned conflict risk assessment report provides a comprehensive understanding of the potential chemical reactions and physical damage risks to goods during transportation. This technical approach effectively identifies potential safety hazards during transport and, through multi-level risk analysis and assessment, ensures the safety of the transportation process, preventing accidents or delays caused by improper cargo combinations or environmental changes.

[0054] Step 3: Extract a list of high-risk cargo combinations from the conflict risk assessment report, and use a dynamic constraint verification mechanism to screen cargo matching schemes to obtain a preliminary loading scheme.

[0055] In one specific embodiment, the process of performing step 3 may specifically include the following steps:

[0056] For a list of high-risk goods combinations, a set of compatible and safe goods combinations is determined by filtering goods matching schemes based on a constraint rule base.

[0057] Based on the cargo combination set, a linear programming algorithm is used to optimize the cargo loading order and generate the first loading plan;

[0058] If the cargo combination in the first loading scheme does not meet the compatibility and safety constraints, the cargo positions are adjusted through a space allocation algorithm to obtain an optimized loading scheme.

[0059] The cargo location and matching information are extracted from the optimized loading plan. A conflict detection algorithm is used to verify whether there is a conflict risk. If not, the optimized loading plan is used as the preliminary loading plan.

[0060] For example, suppose a shipping company is carrying out less-than-container-load (LCL) shipping of various goods, comprising 5 shipments, the attributes of which are shown in Table 1:

[0061] Table 1

[0062]

[0063] Based on the type and characteristics of the goods, the system selects compatible and safe combinations of goods from a constraint rule base. The constraint rule base contains safety requirements and compatibility rules regarding the physical properties, chemical reactivity, temperature stability, weight, and volume of the goods. These rules ensure that interactions between goods during loading do not lead to incompatibility issues, such as chemical reactions or physical damage. Assume the following preset rules: 1) Chemicals A01 and B02 can be combined if their chemical reactivity is below 0.8; 2) Heavy item C03 cannot be combined with other goods and must be loaded separately; 3) Non-hazardous item D04 can be combined with any type of goods, but the loading order must be ensured to avoid damage to other goods. Based on these rules, the system selects the following combinations of goods: 1) Combination 1: A01 and B02; 2) Combination 2: D04 and A01; 3) Combination 3: E05 and D04.

[0064] Based on the selected cargo combination set, the system employs a linear programming algorithm to optimize the cargo loading sequence, ensuring that the loading plan meets both cargo compatibility requirements and maximizes space utilization. The linear programming algorithm optimizes the objective function under constraints, ensuring that the cargo loading sequence meets optimal space utilization and safety requirements. In this process, the objective function can include maximizing loading space utilization, minimizing cargo safety risks, and maximizing transportation efficiency. For example, the objective function can be defined as: Objective Function = Maximize Space Utilization + Minimize Loading Time + Minimize Risk.

[0065] Using a linear programming algorithm, the system optimizes the loading sequence, ensuring that high-density cargo such as C03 and E05 are rationally arranged in the lower part of the container, while lighter chemicals (such as A01 and B02) are arranged in the upper layer. This ensures efficient use of loading space while reducing potential risks. Assume the generated first loading plan is as follows: Loading sequence: A01, B02, D04, E05, C03; Space allocation: C03 is located at the bottom of the container, A01, B02, and E05 are in the upper layer, and D04 is in between.

[0066] If certain cargo combinations in the first loading plan do not meet compatibility and safety constraints (e.g., A01 and E05 may react due to temperature mismatch), adjustments are needed using a space allocation algorithm. This algorithm rearranges the cargo based on its size, weight, and the volume constraints of the loading space, ensuring that each batch of cargo is positioned safely within the container and that the loading space is used optimally. In this case, the system will automatically adjust the cargo positions to ensure that incompatible cargo is not placed together, avoiding conflicts. Assuming the adjusted loading plan is as follows, with the adjustment order: A01, D04, B02, E05, C03; and the adjusted space allocation: E05 and D04 are assigned to different positions, ensuring sufficient space isolation between them.

[0067] From the optimized loading plan, the location and combination information of the goods are extracted and verified using a conflict detection algorithm. The conflict detection algorithm checks whether each pair of goods might cause conflict or safety issues based on characteristics such as the chemical reactivity, physical stability, and temperature of the goods. Assuming the detection result is: no conflict risk is found. Since the conflict detection result is negative, the system uses the optimized loading plan as the preliminary loading plan and ultimately generates a container loading plan.

[0068] By combining this dynamic constraint verification mechanism with optimization algorithms, the loading plan for each batch of goods can be determined efficiently and accurately, ensuring safety and efficiency during transportation. This application, by comprehensively considering the multidimensional characteristics and safety requirements of the goods, avoids accidents caused by potential risks such as mismatched physical properties or chemical reactions, effectively improves the utilization rate of loading space, and reduces risks during transportation.

[0069] Step 4: Obtain time data for ship navigation, port loading and unloading, and customs clearance, construct a real-time time dataset, and generate a comprehensive time limit prediction model based on it.

[0070] In one specific embodiment, the process of performing step 4 may specifically include the following steps:

[0071] We acquire ship navigation data, port loading and unloading data, and customs clearance data to obtain a multi-source time dataset. After cleaning, alignment, and linear interpolation of missing values, we obtain a standardized dataset.

[0072] Feature extraction is performed on the standardized dataset to obtain the time feature dataset. The random forest algorithm is then used to extract the internal time series features to form a time series feature vector.

[0073] A long short-term memory network model is trained using temporal feature vectors to obtain a comprehensive time-limited prediction model.

[0074] Specifically, because data from different sources may contain missing values ​​or inconsistent formats, data cleaning and alignment are necessary. The cleaning process includes removing erroneous or incomplete records, filling in missing values, and ensuring data consistency and integrity. During this process, missing values ​​can be filled using linear interpolation to guarantee data continuity and accuracy. After data cleaning, a standardized dataset is obtained, ensuring that the units of measurement for each time variable are consistent.

[0075] By extracting key time-related features, such as time differences in transportation stages, temporal patterns of each stage, and delay trends, a feature dataset reflecting the dynamic changes in the transportation process is formed. These features contain information about potential delays and schedule changes encountered by goods during transportation, providing effective input for subsequent prediction models. To further improve prediction accuracy, the random forest algorithm is used to analyze these time feature datasets and extract their internal temporal features. The random forest algorithm constructs multiple decision trees to evaluate and optimize each feature, identifying the most predictive temporal features and generating a temporal feature vector.

[0076] LSTM, as an effective time-series data processing model, can capture long-term dependencies through its memory mechanism, demonstrating excellent performance in handling complex time-series data. LSTM continuously adjusts its internal weights and states to optimize predictions of future time limits. A trained LSTM model can predict the time limits of goods during transportation based on real-time data, helping management systems optimize loading plans and adjust transportation arrangements to cope with potential delays and unforeseen circumstances.

[0077] By comprehensively analyzing multi-source time data and combining it with advanced algorithms for prediction, this invention can not only provide more accurate time limit predictions, but also dynamically adjust transportation plans, reduce delays and uncertainties in the transportation process, significantly improve the efficiency and accuracy of shipping management for consolidation containers, and solve the problem of transportation delays caused by the lack of real-time time limit predictions.

[0078] Step 5: Based on the comprehensive time limit prediction model, sort the delivery time requirements of each batch of goods, and use an automatic weight adjustment algorithm to optimize the loading priority to obtain a loading priority list.

[0079] In one specific embodiment, the process of performing step 5 may specifically include the following steps:

[0080] Based on the time constraints and batch information of each batch of goods obtained from the goods identifier, an initial goods list is generated;

[0081] Based on the initial goods list, the delivery time requirements for each batch of goods are ranked to obtain the time limit ranking result;

[0082] If there are time constraint conflicts in the time limit sorting results, the automatic weight adjustment algorithm is called to calculate the loading priority of each batch of goods based on the dynamic weight parameters and generate a priority adjustment list.

[0083] Based on the priority adjustment list, the loading priority is iteratively optimized using an optimization algorithm to obtain the loading priority list.

[0084] Specifically, the time constraints for each batch of goods include the predetermined delivery time requirement and the time requirements for each stage of transportation. Based on these time requirements, the delivery time of each batch of goods is ranked to obtain a time limit ranking result. The ranking process is based on the delivery time of the goods and their transportation priority, ensuring that goods with more urgent delivery time are processed first during transportation.

[0085] If time constraint conflicts occur in the time-limit ranking results, meaning that the delivery deadlines of some goods cannot be met simultaneously, an automatic weight adjustment algorithm is activated. This algorithm adjusts the loading priority of goods based on dynamic weight parameters to ensure that time-limit conflicts are effectively resolved. The dynamic weight parameters are adjusted in real time based on the time-limit constraints, transportation priorities, and loading requirements of each batch of goods, assigning higher priority to more urgent goods. Through this adjustment process, the needs of goods with urgent delivery deadlines can be prioritized while maintaining the stability of the overall transportation plan.

[0086] Based on the adjusted priority list, the system uses an optimization algorithm to iteratively optimize the loading priorities, resulting in a final loading priority list. In each iteration, the optimization algorithm considers loading space, transportation efficiency, and time requirements, adjusting the loading order to maximize space utilization and minimize transportation delays. During each iteration, the loading order is optimized based on priority adjustments, ultimately generating a priority list that balances timeliness and loading efficiency. This list provides clear guidance for subsequent loading and transportation scheduling, ensuring that each batch of goods is processed promptly according to its time requirements, while optimizing the utilization of transportation resources.

[0087] By combining a comprehensive time-limit prediction model with an automatic weight adjustment algorithm, the system can automatically adjust the transportation plan based on real-time data, ensuring that goods can be delivered smoothly within the specified time limit, significantly improving transportation efficiency and reducing the risk of delays caused by improper loading.

[0088] Step 6: Extract the highest priority cargo from the loading priority list and dynamically adjust it in conjunction with the preliminary loading plan. If a priority conflict is detected, the container positions are reassigned to obtain the final loading plan.

[0089] In one specific embodiment, the process of performing step 6 may specifically include the following steps:

[0090] Obtain the set of highest priority goods, and calculate the initial loading positions using a linear programming algorithm based on the container position allocation rules in the preliminary loading plan to obtain the preliminary loading configuration;

[0091] Analyze the initial loading configuration. If there are priority conflicts, recalculate the allocation weights based on the degree of conflict and container capacity constraints to obtain the adjusted loading configuration.

[0092] A greedy algorithm is used to perform secondary optimization on the adjusted loading configuration in order to reallocate container positions and obtain an optimized loading scheme;

[0093] The correspondence between container locations and cargo is extracted from the optimized loading plan. If there is unallocated cargo, it is reallocated according to the remaining container capacity to obtain a supplementary loading plan.

[0094] If the supplementary loading plan meets the stability and priority requirements, it will be determined as the final loading plan.

[0095] Specifically, the highest priority cargo is retrieved from the loading priority list, and a preset priority sorting algorithm is used to determine the cargo identifier and priority value, resulting in a set of highest priority cargoes. Subsequently, by analyzing the time constraints and transportation requirements of these cargoes, and combining the container locations and other constraints already allocated in the preliminary loading plan, the cargoes are dynamically adjusted.

[0096] For example, container location allocation rules include prioritizing the loading of heavy cargo in the bottom area of ​​the container and prioritizing the loading of light cargo in the upper area of ​​the container.

[0097] The loading location of each cargo is influenced by multiple constraints, including container capacity and the dimensions and weight of each cargo. During dynamic adjustment, a linear programming algorithm is used to calculate the initial loading location for each batch of cargo, based on the container's space limitations and the type and characteristics of the cargo. The linear programming algorithm calculates the optimal location for each cargo by setting an objective function (e.g., maximizing space utilization) and combining it with container space constraints (e.g., bottom area space, top area space, and cargo dimensions and weight). In this process, the algorithm not only considers the loading order but also ensures compatibility and safety among cargoes during loading. For example, assuming a container with a capacity of 10m³, a bottom area that can hold 6m³ of cargo, and a top area that can hold 4m³ of cargo, if the highest priority cargo (e.g., heavy cargo A01 and light cargo B02) is extracted from the loading priority list, an initial location will be assigned to each cargo according to the linear programming algorithm. Heavy cargo A01 will be preferentially assigned to the bottom area, occupying 3m³ of space, while light cargo B02 will be assigned to the top area, occupying 2m³ of space.

[0098] If priority conflicts exist in the initial loading configuration—for example, if the loading positions of some high-priority goods conflict with those of other goods, or if the combination of goods does not meet safety constraints—the system will recalculate the allocation weights and adjust the loading positions of the goods based on the degree of conflict and container capacity constraints. In this case, conflicting goods will be reconsidered and rearranged to ensure that high-priority goods are loaded first, without affecting the safety and loading efficiency of other goods.

[0099] The greedy algorithm maximizes space utilization while considering cargo safety and loading order by progressively optimizing the cargo layout within the container. In each optimization, the algorithm selects the current optimal solution and gradually adjusts the cargo positions within the container until the entire loading plan meets all safety and space utilization requirements. If there are unallocated cargoes, their positions are reallocated based on the remaining container capacity, generating supplementary loading plans. These supplementary loading plans ensure that every piece of cargo is allocated appropriately and does not violate capacity and safety constraints. Subsequently, the supplementary loading plans are checked again to ensure they meet stability and priority requirements; if a plan meets these requirements, it is selected as the final loading plan.

[0100] By dynamically adjusting loading plans and optimizing loading priorities, it is possible to better balance transportation timeliness and safety, ensure that each batch of goods is handled reasonably according to priority, optimize the use of container space, and improve overall transportation efficiency and safety.

[0101] Step 7: Monitor the ship's navigation and port operations in real time. If a delay risk is detected, trigger the emergency procedure to adjust the loading priority or reallocate containers and obtain the adjusted loading plan.

[0102] In one specific embodiment, the process of performing step 7 may specifically include the following steps:

[0103] Acquire ship navigation status and port operation status data to generate a real-time status dataset;

[0104] If the real-time status dataset indicates that the delay risk exceeds the fourth preset value, the random forest algorithm is used to analyze the risk factors and determine the delay risk level.

[0105] Based on the delay risk level, appropriate rules are extracted from the dynamic scheduling rule base to generate a priority adjustment strategy;

[0106] Based on the priority adjustment strategy, a linear programming algorithm is used to optimize the container allocation scheme, resulting in a preliminary adjustment scheme.

[0107] The initial adjustment plan is verified through simulation analysis. It is then determined whether the verification results meet the preset efficiency threshold. If so, the initial adjustment plan is adopted as the adjusted loading plan. If not, the real-time status dataset is re-acquired, the container allocation plan is iteratively optimized, and the adjusted loading plan is obtained.

[0108] Specifically, real-time data on vessel navigation status and port operations status is collected through sensors, GPS devices, and port operation management systems. The real-time status dataset includes information such as the vessel's current position, speed, estimated arrival time, port operation progress, and loading / unloading speed.

[0109] If the real-time status dataset shows that the delay risk exceeds a fourth preset value (e.g., the scheduled arrival time is delayed beyond a certain time threshold), the random forest algorithm is used to analyze risk factors and assess the delay risk level. The random forest algorithm trains multiple decision trees to analyze real-time status data, identifying key factors that may cause delays, such as port loading and unloading progress, ship speed, and weather conditions. Based on this, the delay risk can be accurately assessed, and a corresponding delay risk level can be generated, reflecting potential problems in the current transportation process.

[0110] The dynamic scheduling rule base contains scheduling schemes to handle different risk levels. For example, in high-risk situations, it prioritizes time-sensitive goods, or in lower-risk situations, it adjusts the loading order of non-urgent goods. Based on these rules, the loading priority of goods is adjusted to minimize delays and ensure the timely delivery of critical goods. Once the priority adjustment strategy is generated, a linear programming algorithm is used to obtain an initial adjustment plan. The linear programming algorithm optimizes the loading order of goods within the container by considering multiple factors such as container capacity, cargo priority, and space constraints, ensuring that the adjusted loading plan meets the new priority requirements and space utilization requirements.

[0111] The efficiency threshold refers to whether the loading, transportation process, and arrival time of goods meet the predetermined requirements within a given time frame. If the initial adjustment plan meets the threshold, it is used as the emergency adjustment loading plan; if it does not meet the threshold, the system will re-acquire the real-time status dataset and iteratively optimize the container allocation plan to ensure that the emergency adjustment loading plan can meet the transportation needs.

[0112] This dynamic adjustment process enables rapid response and flexible adjustment of transportation plans based on real-time data feedback during actual transportation, reducing delays and improving transportation efficiency. This intelligent loading optimization scheme, based on delay risk analysis and real-time adjustment, effectively solves the problem of lack of real-time monitoring and dynamic adjustment in traditional methods, ensuring on-time delivery of goods, improving the utilization efficiency of container space, and optimizing overall logistics scheduling.

[0113] Step 8: Extract the adjustment records from the adjusted loading plan, verify them in conjunction with the real-time time dataset, determine whether the time limit requirements of all goods are met, and obtain the verification results.

[0114] In one specific embodiment, the process of performing step 8 may specifically include the following steps:

[0115] Extract records containing cargo information and scheduling instructions from the adjustment records to obtain a structured adjustment dataset;

[0116] The structured adjustment dataset is compared with the real-time dataset by timestamp. If there are goods that do not meet the time limit requirements, the linear programming algorithm is called to optimize the loading plan after emergency adjustment and the scheduling instructions are rearranged to obtain the optimized adjustment dataset.

[0117] Extract the updated scheduling instructions and cross-validate the time limits of all goods with the real-time time dataset to obtain the validation dataset;

[0118] Based on the verification dataset, a decision tree algorithm is used to determine whether there are goods that do not meet the time limit requirements. If so, the scheduling instructions are updated iteratively, and the verification is performed again by combining the real-time time dataset to obtain the verification result.

[0119] Specifically, the structured adjustment dataset includes information such as the loading sequence, loading location, and scheduling instructions for each batch of goods, which can accurately describe the allocation of goods in containers and their relationship with time limit requirements.

[0120] The generated structured adjustment dataset is compared with the real-time acquired time dataset, particularly comparing the loading plan and actual transportation time for each batch of goods based on timestamps. If the comparison results show that the time requirements for some goods cannot be met, the optimization process is triggered. A linear programming algorithm is invoked to optimize the emergency-adjusted loading plan, readjusting the position and loading order of the goods. The linear programming algorithm calculates a new optimal loading plan based on factors such as container capacity, cargo priority, and time constraints, and generates an updated adjustment dataset.

[0121] Updated scheduling instructions are extracted from the optimized adjustment dataset and cross-validated with the real-time dataset to ensure on-time delivery for all goods. Cross-validation compares the loading time, estimated arrival time, and actual transport time of each batch of goods to determine whether on-time delivery is possible. The validation dataset compiles the loading status, transport progress, and time limits for all goods. Decision trees provide fast, transparent, and flexible decision support when validating time limits, making them particularly suitable for handling multi-dimensional and multi-constraint loading and transport scheduling problems. Based on the validation dataset, the decision tree algorithm automatically determines whether there are unqualified loading schemes based on factors such as the time limits of the goods, transport status, and loading order. If the decision tree algorithm detects time limit conflicts or some goods fail to be delivered on time, it returns to the initial loading scheme, iterates and updates the scheduling instructions according to the new time constraints, and re-validates them using the real-time dataset until a final loading scheme that meets the time limits is obtained.

[0122] By combining linear programming, decision trees, and real-time data verification, it can flexibly respond to changes in the transportation process, ensure that each batch of goods can be delivered smoothly within the specified time, and avoid delays or waste of resources due to improper loading, thereby improving the flexibility and responsiveness of the entire logistics chain.

[0123] Step 9: Based on the verification results, generate a container loading plan and scheduling report.

[0124] In one specific embodiment, the process of performing step 9 may specifically include the following steps:

[0125] Container parameters and transportation time limits are read from the logistics management system to form a structured dataset;

[0126] An initial loading plan is generated based on a structured dataset using a linear programming algorithm.

[0127] Determine whether the initial loading plan meets the container capacity and weight constraints. If it does, use it as the container loading plan. If not, adjust the parameter weights until an initial loading plan that meets the container capacity and weight constraints is obtained.

[0128] Based on the container loading plan and transportation time limit, a time-limited scheduling plan is generated using a time window algorithm;

[0129] Determine whether the time-limited scheduling plan meets the time-limit control threshold. If it does, generate a scheduling report accordingly. If not, optimize the weights of the time window algorithm parameters until a time-limited scheduling plan that meets the requirements is generated.

[0130] Specifically, the structured dataset includes information such as container capacity, time limits for each shipment, cargo type, weight, and volume. Based on this data, a preliminary loading plan is generated using a linear programming algorithm. In this process, by setting an objective function and constraints, and considering factors such as container space and weight limitations, as well as cargo priority and time limits, the initial loading scheme is calculated.

[0131] For example, the initial loading plan is as follows: A01: volume 3m³, weight 2000kg; B02: volume 2m³, weight 1500kg; C03: volume 4m³, weight 3000kg; D04: volume 1.5m³, weight 1000kg; E05: volume 2.5m³, weight 1800kg. At this point, the initial calculation of the total container capacity and total weight is: Total capacity = 3m³ + 2m³ + 4m³ + 1.5m³ + 2.5m³ = 13m³, Total weight = 2000kg + 1500kg + 3000kg + 1000kg + 1800kg = 9300kg. Since the maximum container capacity is limited to 10m³, and the total capacity is 13m³, this exceeds the volume limit. Simultaneously, the maximum load capacity of the container is 8000kg, and the total weight is 9300kg, which also exceeds the weight limit. Therefore, the initial loading plan did not meet the container's capacity and weight constraints.

[0132] To meet the constraints, the system adjusts the parameter weights and re-optimizes until an effective loading plan that meets the container's capacity and weight requirements is obtained. For example, this might involve reducing the loading of some low-priority goods or adjusting their positions. Assuming the system, after adjusting according to linear programming, chooses to remove C03 (volume 4m³, weight 3000kg), the new loading plan is: A01: volume 3m³, weight 2000kg; B02: volume 2m³, weight 1500kg; D04: volume 1.5m³, weight 1000kg; E05: volume 2.5m³, weight 1800kg. At this point: Total capacity = 3m³ + 2m³ + 1.5m³ + 2.5m³ = 9m³; Total weight = 2000kg + 1500kg + 1000kg + 1800kg = 6300kg. The new loading plan meets the container's capacity and weight constraints, therefore it will be used as the container loading plan.

[0133] Based on the above container loading plan, the system also needs to ensure timely delivery of goods. Assume the following transportation time limits: A01 transportation time limit 48 hours; B02 transportation time limit 36 ​​hours; D04 transportation time limit 24 hours; E05 transportation time limit 60 hours. The time window algorithm generates a time-limited scheduling plan based on the time limit requirements of each cargo and the time constraints of each transportation link. Assume that, based on the actual transportation schedule, the transportation time limits for B02 and D04 are more pressing, therefore the system needs to prioritize these cargoes. Based on the priority, the system adjusts the loading order and determines the shortest transportation time and feasible transportation plan for each batch of cargo. Assume the calculated time-limited scheduling plan is: B02: priority loading, transportation time limit 36 ​​hours; D04: second priority, transportation time limit 24 hours; A01: second-second priority, transportation time limit 48 hours; E05: last loading, transportation time limit 60 hours. The system formulates a time-limited scheduling plan based on these priorities and time limits.

[0134] Assume the time limit control threshold is 48 hours, meaning the delivery time for each batch of goods cannot exceed 48 hours. In this case, B02 and D04 meet the time limit requirements, but E05's delivery time is 60 hours, exceeding the time limit control threshold. The system detects a time limit conflict and triggers further optimization. To resolve the time limit conflict, the system adjusts the parameter weights of the time window algorithm, for example, by adjusting priorities or reallocating the loading order, ensuring that E05's time limit requirement is not violated. Assume that by optimizing and adjusting the loading order and transportation route of E05, the system ultimately obtains a new time-limit scheduling plan, ensuring that all goods can be transported within the specified time limit.

[0135] The system generates a scheduling report based on the new scheduling plan. The report includes information such as the loading sequence, transportation time, estimated arrival time, and whether the time limit requirements are met for each batch of goods. Through a data transmission protocol, the loading plan and scheduling report are integrated into a structured output and transmitted to the logistics management system to obtain the final plan recorded by the system. For the final plan recorded by the system, a log recording mechanism is used to store the operation history, generating traceable business execution records.

[0136] The time window algorithm is an algorithm used for scheduling and optimization problems, especially suitable for solving time-constrained task scheduling problems. In the logistics and transportation fields, the time window algorithm is often used to optimize the loading, transportation, and distribution of goods to ensure that tasks are completed within a predetermined time frame while maximizing resource utilization. Its core lies in the concept of a "time window." For each task (such as transporting goods or performing loading and unloading operations), there is a specific time window, indicating that the task must start or be completed within this time frame. Specifically, each task has a range of start and end times, i.e., the time window. If a task cannot be completed within the specified time window, it is considered a breach of contract, which may lead to delays, penalties, or other problems. Common time window algorithms include the Greedy Algorithm, Dynamic Programming, Genetic Algorithm (GA), Simulated Annealing (SA), and Particle Swarm Optimization (PSO).

[0137] By combining linear programming and time window algorithms, the system can ensure timely delivery of goods while maintaining the utilization rate of container space, thereby improving transportation efficiency and reliability. This solves problems such as inaccurate loading and scheduling, resource waste, and transportation delays in traditional logistics management.

[0138] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for managing consolidation container shipping based on an Internet of Things (IoT) platform, characterized in that, The method includes: Step 1: Obtain the goods list, extract the chemical and physical properties data of each batch of goods to form an initial attribute dataset, and use a classification database to classify the goods to obtain the attribute feature set of the goods. Step 2: Analyze the interaction between chemical properties and physical characteristics based on the attribute feature set. If a potential chemical reaction or physical damage is detected, a conflict risk assessment report is generated. Step 3: Extract a list of high-risk cargo combinations from the conflict risk assessment report, and use a dynamic constraint verification mechanism to screen cargo matching schemes to obtain a preliminary loading scheme; Step 4: Obtain time data for ship navigation, port loading and unloading, and customs clearance, construct a real-time time dataset, and generate a comprehensive time limit prediction model based on it. Step 5: Based on the comprehensive time limit prediction model, sort the delivery time limit requirements for each batch of goods, and use an automatic weight adjustment algorithm to optimize the loading priority to obtain a loading priority list; Step 6: Extract the highest priority cargo from the loading priority list and dynamically adjust it in conjunction with the preliminary loading plan. If a priority conflict is detected, the container positions are reassigned to obtain the final loading plan. Step 7: Monitor the ship's navigation and port operation status in real time. If a delay risk is detected, trigger the emergency procedure to adjust the loading priority or reallocate containers and obtain the adjusted loading plan. Step 8: Extract the adjustment records from the adjusted loading plan, verify them in conjunction with the real-time time dataset, determine whether the time limit requirements of all goods are met, and obtain the verification results; Step 9: Based on the verification results, generate a container loading plan and scheduling report; Step 3 includes: For the list of high-risk goods combinations, goods matching schemes are screened based on a constraint rule base to determine a compatible and safe set of goods combinations; Based on the cargo combination set, a linear programming algorithm is used to optimize the cargo loading order and generate a first loading plan. If the cargo combination in the first loading scheme does not meet the compatibility and safety constraints, the cargo position is adjusted through a space allocation algorithm to obtain an optimized loading scheme. The cargo location and matching information are extracted from the optimized loading plan, and a conflict detection algorithm is used to verify whether there is a conflict risk. If not, the optimized loading plan is used as the preliminary loading plan.

2. The method according to claim 1, characterized in that, Step 1 includes: Based on the classification database, the initial attribute dataset is matched, and the K-nearest neighbor algorithm is used to determine the preliminary classification of each batch of goods, resulting in a classification result set. For the classification result set, principal component analysis algorithm is used to extract the main features of chemical properties and physical characteristics to form a simplified feature set; Determine whether the matching degree between the simplified feature set and the preset rules in the classification database is higher than a first preset value. If yes, determine the attribute feature set through attribute analysis. If no, re-extract data from the initial attribute dataset to obtain an updated attribute dataset. Based on the updated attribute dataset, the support vector machine algorithm is used to perform secondary classification of the goods batch to obtain an optimized classification result set. Based on the correlation analysis between the optimized classification result set and the simplified feature set, the attribute feature set of each batch of goods is determined, and the final feature description set is obtained. Based on the final feature description set, a classification label and attribute feature mapping for each batch of goods are generated, and the attribute feature set is obtained.

3. The method according to claim 1, characterized in that, Step 2 includes: Obtain the chemical composition and physical state data of the cargo from the attribute feature set; If the reactivity of the chemical component exceeds the second preset value, the chemical reaction path is calculated using a thermodynamic model, and the reaction risk level is determined. Based on the aforementioned reaction risk level, the corresponding temperature threshold and pressure conditions are determined, and finite element analysis is used to simulate changes in physical properties to obtain stress distribution results. If the stress distribution results show that the mechanical strength is lower than the third preset value, then the risk of physical damage is determined by analyzing the corrosion tendency and thermal stability through molecular dynamics simulation. Based on the analysis results of the corrosion tendency and thermal stability, key features are extracted, and support vector machine is used to classify the risk level to obtain the classified risk assessment results. Based on the risk assessment results, density distribution and surface tension data are obtained, and the comprehensive impact intensity of each risk factor is calculated using the numerical integration method, thereby determining the conflict risk level. Based on the conflict risk level, a conflict risk assessment report is generated that includes the risks of chemical reaction and physical damage.

4. The method according to claim 1, characterized in that, Step 4 includes: We acquire ship navigation data, port loading and unloading data, and customs clearance data to obtain a multi-source time dataset. After cleaning, alignment, and linear interpolation of missing values, we obtain a standardized dataset. Feature extraction is performed on the standardized dataset to obtain a time feature dataset. The random forest algorithm is then used to extract the time-series features within the dataset to form a time-series feature vector. The long short-term memory network model is trained using the time-series feature vectors to obtain the comprehensive time-limit prediction model.

5. The method according to claim 1, characterized in that, Step 5 includes: Based on the time constraints and batch information of each batch of goods obtained from the goods identifier, an initial goods list is generated; Based on the initial goods list, the delivery time requirements for each batch of goods are arranged to obtain a time limit ranking result; If there is a time constraint conflict in the time limit sorting results, the automatic weight adjustment algorithm is invoked to calculate the loading priority of each batch of goods based on the dynamic weight parameters and generate a priority adjustment list. Based on the priority adjustment list, the loading priority is iteratively optimized using an optimization algorithm to obtain the loading priority list.

6. The method according to claim 1, characterized in that, Step 6 includes: Obtain the highest priority cargo set, and calculate the initial loading position using a linear programming algorithm based on the container position allocation rules in the preliminary loading plan to obtain the preliminary loading configuration; If priority conflicts exist in the initial loading configuration, the weights are recalculated based on the degree of conflict and container capacity constraints to obtain the adjusted loading configuration. A greedy algorithm is used to perform secondary optimization on the adjusted loading configuration to reallocate container positions and obtain an optimized loading scheme. The correspondence between container locations and cargo is extracted from the optimized loading scheme. If there is unallocated cargo, it is reallocated according to the remaining container capacity to obtain a supplementary loading scheme. If the supplementary loading scheme meets the stability and priority requirements, it will be determined as the final loading scheme.

7. The method according to claim 1, characterized in that, Step 7 includes: Acquire ship navigation status and port operation status data to generate a real-time status dataset; If the real-time status dataset indicates that the delay risk exceeds the fourth preset value, then the random forest algorithm is used to analyze the risk factors and determine the delay risk level. Based on the delay risk level, appropriate rules are extracted from the dynamic scheduling rule base to generate a priority adjustment strategy; Based on the priority adjustment strategy, a linear programming algorithm is used to optimize the container allocation scheme to obtain a preliminary adjustment scheme. The preliminary adjustment scheme is verified by simulation analysis. It is determined whether the verification result meets the preset efficiency threshold. If so, the preliminary adjustment scheme is adopted as the adjusted loading scheme. If not, the real-time status dataset is re-acquired, the container allocation scheme is iteratively optimized, and the adjusted loading scheme is obtained.

8. The method according to claim 1, characterized in that, Step 8 includes: Extract records containing cargo information and scheduling instructions from the adjustment records to obtain a structured adjustment dataset; The structured adjustment dataset is compared with the real-time dataset by timestamp. If there are goods that do not meet the time limit requirements, the linear programming algorithm is called to optimize the adjusted loading scheme and the scheduling instructions are rearranged to obtain the optimized adjustment dataset. Extract the updated scheduling instructions and cross-validate the time limits of all goods with the real-time time dataset to obtain the validation dataset; Based on the verification dataset, a decision tree algorithm is used to determine whether there are any goods that do not meet the time limit requirements. If so, the scheduling instructions are updated iteratively, and the verification is performed again in conjunction with the real-time time dataset to obtain the verification result.

9. The method according to claim 1, characterized in that, Step 9 includes: Container parameters and transportation time limits are read from the logistics management system to form a structured dataset; Based on the structured dataset, an initial loading plan is generated using a linear programming algorithm; Determine whether the initial loading plan meets the container capacity and weight constraints. If it does, use it as the container loading plan. If not, adjust the parameter weights until an initial loading plan that meets the container capacity and weight constraints is obtained. Based on the container loading plan and the transportation time limit, a time-limited scheduling plan is generated using a time window algorithm; Determine whether the time-limited scheduling plan meets the time-limited control threshold. If it does, generate the scheduling report accordingly. If not, optimize the weights of the time window algorithm parameters until a time-limited scheduling plan that meets the requirements is generated.

Images