Intelligent stowage method and system

Abstract

The application discloses the technical field of intelligent stowage and discloses an intelligent stowage method and system, which comprises the following steps: based on the basic information of goods and stowage, using principal component analysis and data cleaning method, feature extraction and data preprocessing are carried out, the characteristics and demand of goods and the conditions and restrictions of stowage are analyzed, and the goods stowage feature data set is generated. In the application, through the application of principal component analysis and data cleaning method, the accuracy and adaptability of the complex relationship between goods and stowage are improved, so that the stowage system can more flexibly cope with the changes of market and operation conditions. Through fine management of the characteristics of goods and the conditions of stowage, more efficient goods-stowage matching is realized, so as to maximize the resource utilization efficiency. By using the contract design algorithm and mechanism design principle, as well as the dynamic programming algorithm and time sensitivity weighting method, appropriate weight is allocated to goods with different time sensitivity and the priority is dynamically adjusted.

Description

Technical Field

[0001] This invention relates to the field of intelligent cabin rearrangement technology, specifically to an intelligent cabin rearrangement method and system. Background Technology

[0002] The field of intelligent cargo scheduling technology involves using algorithms and computer systems to intelligently arrange cargo placement based on various order information, thereby improving loading efficiency, optimizing space utilization, and reducing transportation costs. This field combines research findings from multiple disciplines, including logistics management, computer science, and artificial intelligence, aiming to solve the problems of inefficiency, high error rates, and insufficient resource utilization inherent in traditional manual cargo scheduling through automation and intelligent methods. Intelligent cargo scheduling systems can process complex data, such as cargo dimensions, weight, destination, and special handling requirements, to develop optimal loading plans.

[0003] The core of intelligent cargo scheduling lies in automatically calculating the optimal cargo loading plan by analyzing and processing order information. Its aim is to maximize cargo space utilization, reduce transportation costs, and ensure the safe and timely arrival of goods at their destination. The method seeks to achieve effects including increased loading speed, optimized ship space utilization, reduced human error, and improved overall logistics efficiency. By achieving these goals, companies can provide more efficient and economical logistics services, thereby gaining a competitive edge in the market.

[0004] While existing technologies have achieved the goals of improving loading efficiency and optimizing space utilization in the field of intelligent cargo scheduling, they are insufficient in terms of flexibility and adaptability in handling the complex relationship between cargo and cargo space, making it difficult to effectively cope with rapid changes in market and operational conditions. Furthermore, the mechanisms for differentiating and prioritizing cargo with varying time sensitivity are inadequate, limiting the rapid response capability for urgent cargo. Regarding the ability to identify and handle abnormal patterns, the lack of full utilization of historical scheduling data for deep learning results in untimely or ineffective handling of abnormal situations. The methods for predicting and managing cargo damage risks during loading and transportation are also insufficient, lacking effective risk assessment and control strategies, which affects the optimization of loading plans and the reduction of losses. Finally, the comprehensive management capability in the face of uncertainty is also insufficient, lacking effective mechanisms to cope with uncertainties such as cargo arrival delays and sudden peak periods, reducing the flexibility and accuracy of scheduling decisions.

[0005] Based on this, the present invention designs an intelligent cabin arrangement method to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent cargo scheduling method to address the shortcomings mentioned in the background art. While existing technologies in the field of intelligent cargo scheduling have achieved the goals of improving loading efficiency and optimizing space utilization, they are insufficient in terms of flexibility and adaptability in handling the complex relationship between cargo and cargo space, making it difficult to effectively cope with rapid changes in market and operational conditions. Furthermore, the mechanisms for differentiating and prioritizing cargo with varying time sensitivity are inadequate, limiting the rapid response capability for urgent cargo. Regarding the identification and handling of abnormal patterns, the lack of full utilization of historical scheduling data for deep learning results in untimely or ineffective handling of abnormal situations. The methods for predicting and managing cargo loss risks during loading and transportation are also insufficient, lacking effective risk assessment and control strategies, which affects the optimization of loading plans and the reduction of losses. Finally, the comprehensive management capability in the face of uncertainty is also insufficient, lacking effective mechanisms to deal with uncertainties such as cargo arrival delays and sudden peak periods, reducing the flexibility and accuracy of scheduling decisions.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an intelligent cabin sorting method, comprising the following steps:

[0008] S1: Based on the basic information of cargo and cargo space, principal component analysis and data cleaning methods are used to extract features and preprocess data, analyze cargo characteristics and demand, as well as cargo space conditions and constraints, and generate cargo space feature dataset.

[0009] S2: Based on the cargo space feature dataset, a cargo-space contract model is designed using contract design algorithms and mechanism design principles. The efficiency and fairness of the model are evaluated, and an optimized contract model set is generated.

[0010] S3: Based on the optimized contract model set, a dynamic programming algorithm and a time-sensitive weighted method are used to allocate weights to the cargo, dynamically adjust the cargo scheduling priority, and generate a time-priority scheduling scheme.

[0011] S4: Based on the time-priority cabin allocation scheme, cluster analysis and isolated forest algorithm are used to analyze historical cabin allocation data and the current cabin allocation scheme, identify abnormal patterns, optimize the cabin allocation process, and generate anomaly identification and adjustment scheme;

[0012] S5: Based on the aforementioned anomaly identification and adjustment scheme, a decision tree model is used to assess the risk of cargo loss during loading and transportation, and risk factors are introduced into the cargo allocation decision to generate risk assessment results.

[0013] S6: Based on the risk assessment results, a probability model is constructed using a Bayesian network, taking into account the uncertainties of cargo arrival delays and sudden peak periods, to provide a basis for scheduling decisions and generate a risk factor management plan;

[0014] S7: Based on the aforementioned risk factor management scheme, a real-time data monitoring method and a random forest algorithm are used to collect key performance indicators during the cabin allocation process in real time, dynamically adjust the cabin allocation strategy, form a closed-loop control, and generate a real-time optimized cabin allocation strategy.

[0015] Preferably, the cargo space feature dataset includes adjusted cargo volume parameters, weight parameters, sensitivity levels, space environmental adaptability parameters, and space configuration; the optimized contract model set includes cost-effectiveness ratios, contract default risk assessments, and time efficiency indicators customized for each cargo space combination; the time-priority scheduling scheme includes time-sensitivity-weighted cargo priority ranking, estimated arrival time, and emergency response level; the anomaly identification and adjustment scheme includes classification of abnormal scheduling events, preventive measures, and timeliness adjustment plans; the risk assessment results include cargo loss probability assessment, potential delay risk index, and safety measures; the risk factor management scheme includes uncertainty response strategies, risk mitigation measures, and emergency preparedness levels; and the real-time optimized scheduling strategy includes adjusted loading order, optimized space allocation schemes, and real-time monitoring and feedback mechanisms.

[0016] Preferably, based on the basic information of cargo and cargo space, the specific steps for generating a cargo space feature dataset are as follows: using principal component analysis and data cleaning methods, feature extraction and data preprocessing are performed, cargo characteristics and demand are analyzed, as well as cargo space conditions and limitations are analyzed.

[0017] S101: Based on the basic information of cargo and cargo space, conduct data quality assessment, use the isnull() and drop_duplicates() methods to check and remove missing values ​​and duplicate records, apply the Z-score normalization method to identify and handle outliers, and generate a clean dataset;

[0018] S102: Based on the clean dataset, perform principal component analysis, use the PCA(n_components=0.95) command to select principal components, retain 95% of the data variability, reduce data dimensionality and retain key variables, and generate a feature set with simplified dimensionality;

[0019] S103: Based on the simplified feature set of the aforementioned dimensions, perform feature extraction, filter key parameters, including cargo volume, weight, type, as well as cargo bay size, location, and environmental conditions, to generate a cargo bay feature dataset.

[0020] Preferably, based on the cargo space feature dataset, a cargo-space contract model is designed using contract design algorithms and mechanism design principles, and the efficiency and fairness of the model are evaluated to generate an optimized contract model set. The specific steps are as follows:

[0021] S201: Based on the cargo space feature dataset, define an initial model for contract design, use the LinearRegression() method to simulate basic terms, including transportation costs and time constraints, and generate an initial contract model framework.

[0022] S202: Based on the initialized contract model framework, apply the mechanism design principle, use the GameTheoryOptimizer() method to optimize the model, adjust the parameters including efficiency_weight=0.5 and fairness_weight=0.5, and generate the optimized contract model;

[0023] S203: Based on the optimized contract model, a comprehensive evaluation is performed. The evaluate_model(performance_metrics=[efficiency, fairness]) method is used to select the cargo-berth contract model set and generate an optimized contract model set.

[0024] Preferably, based on the optimized contract model set, the specific steps for allocating weights to the cargo using dynamic programming algorithm and time-sensitive weighted method, dynamically adjusting the cargo's scheduling priority, and generating a time-priority scheduling scheme are as follows:

[0025] S301: Based on the optimized contract model set, initialize the dynamic programming environment, set the state space as the priority of each type of cargo for berthing, and the action space as the berthing allocation decision. Use the initialize_dp_environment(states, actions) function to initialize and generate the dynamic programming initialization environment.

[0026] S302: Based on the dynamic programming initialization environment, apply the time-sensitive weighting method to assign weights to each type of goods using the assign_weights(data, time_sensitivity) function, where data is the characteristics of the goods and the time_sensitivity parameter is set according to the urgency of the goods to generate a time-sensitive weight allocation table.

[0027] S303: Based on the time-sensitive weight allocation table, execute the dynamic programming algorithm, and dynamically adjust the cargo scheduling priority through the dynamic_programming(environment, weights) function, giving priority to time-sensitive cargo, and generating a time-priority scheduling scheme.

[0028] Preferably, based on the time-priority cabin allocation scheme, the specific steps for analyzing historical cabin allocation data and the current cabin allocation scheme using cluster analysis and the isolated forest algorithm, identifying abnormal patterns, optimizing the cabin allocation process, and generating anomaly identification and adjustment schemes are as follows:

[0029] S401: Based on the time-priority cabin allocation scheme, the historical cabin allocation data is grouped using cluster analysis. The KMeans(n_clusters=3) function is used to select the number of clusters according to the data characteristics, identify three key cabin allocation patterns, and generate cabin allocation pattern clustering results.

[0030] S402: Based on the clustering results of the cabin allocation pattern, apply the IsolationForest (n_estimators=100, contamination=auto) algorithm to perform anomaly detection of the current cabin allocation scheme, identify abnormal cabin allocation events that differ from historical patterns, and generate basic anomaly identification results.

[0031] S403: Based on the aforementioned basic anomaly identification results, optimize and adjust the cabin scheduling process by using the optimize_scheduling(report, constraints) method, formulate adjustment strategies, and generate anomaly identification and adjustment schemes.

[0032] Preferably, based on the anomaly identification and adjustment scheme, the specific steps for using a decision tree model to assess the loss risk of goods during loading and transportation, and introducing risk factors into the cargo allocation decision to generate risk assessment results are as follows:

[0033] S501: Based on the anomaly identification and adjustment scheme, collect and organize data features associated with loss risk, including organizing cargo type, loading conditions and historical loss records from the anomaly identification and adjustment scheme, creating a data frame using the pandas.DataFrame method, and generating a risk assessment dataset;

[0034] S502: Based on the aforementioned risk assessment dataset, a decision tree model is trained using DecisionTreeClassifier(criterion=gini, max_depth=4) to generate a loss risk assessment model.

[0035] S503: Based on the aforementioned loss risk assessment model, the model.predict(features) method is used to assess the loss risk of each cargo in the current cargo allocation plan and generate risk assessment results.

[0036] Preferably, based on the risk assessment results, a probability model is constructed using a Bayesian network, taking into account uncertainties such as cargo arrival delays and sudden peak periods, to provide a basis for scheduling decisions and generate a risk factor management plan. The specific steps are as follows:

[0037] S601: Based on the risk assessment results, define the nodes and edges of the Bayesian network, build the model using the BayesianModel method of the pgmpy library, the nodes include the risk of delayed arrival of goods, the risk of sudden peak period, and the risk of loss of goods, and set the dependencies between the nodes to generate the Bayesian network structure.

[0038] S602: Based on the Bayesian network structure, the model is trained by a Bayesian estimator. The fit() method is used, with the input parameter being the dataset of the risk assessment results. The conditional probability distribution between nodes is calculated to generate the Bayesian network after parameter training.

[0039] S603: Based on the Bayesian network learned from the parameters, execute the probabilistic inference query() method to evaluate the impact of cargo arrival delays and sudden peak periods on loss risk, and generate a risk factor management plan.

[0040] Preferably, based on the aforementioned risk factor management scheme, the specific steps for using real-time data monitoring methods and random forest algorithms to collect key performance indicators in real time during the cabin allocation process, dynamically adjust the cabin allocation strategy, form closed-loop control, and generate a real-time optimized cabin allocation strategy are as follows:

[0041] S701: Based on the aforementioned risk factor management scheme, real-time data collection is performed. The sensor_data=collect_real_time_data(sensors) function is used to collect data on cabin utilization and loading speed, and a real-time monitoring dataset is generated.

[0042] S702: Based on the real-time monitoring dataset, the RandomForestClassifier(n_estimators=100, max_features=auto, random_state=42) algorithm is used to analyze the data, identify the abnormal cabin arrangement patterns, and generate abnormal efficiency pattern identification results.

[0043] S703: Based on the efficiency anomaly pattern recognition results, dynamically adjust the cabin scheduling strategy, use the adjust_scheduling(strategy, feedback) function to adjust the strategy, apply the simulated annealing algorithm to optimize the adjusted cabin scheduling strategy, and generate a real-time optimized cabin scheduling strategy.

[0044] An intelligent cabin allocation system includes a data quality assessment module, a feature engineering and optimization module, a contract model design module, a cabin allocation priority adjustment module, an anomaly pattern recognition module, a risk assessment module, a risk management strategy module, and a real-time optimization feedback module.

[0045] The data quality assessment module uses basic information about cargo and cargo space to clean and preprocess data by employing methods such as null value checking and deduplication, as well as outlier handling techniques, to generate a clean dataset.

[0046] The feature engineering and optimization module is based on a clean dataset. Through principal component analysis and feature selection techniques, it reduces data dimensionality while retaining key variables, extracts key features of cargo and cargo space, and generates a cargo space feature dataset.

[0047] The contract model design module is based on a cargo hold feature dataset. It uses linear regression to simulate basic terms and combines the Nash equilibrium solution method to balance efficiency and fairness, thereby optimizing the contract model and generating an optimized contract model set.

[0048] The cargo allocation priority adjustment module is based on an optimized contract model set. It applies dynamic programming and time-sensitive weighted methods to allocate priorities to cargo, dynamically adjusts the cargo allocation order, and generates a time-priority cargo allocation scheme.

[0049] The abnormal pattern recognition module is based on the time-first cabin scheduling scheme. It uses cluster analysis and the isolated forest algorithm to identify abnormal patterns in the cabin scheduling data, optimizes the cabin scheduling process, and generates anomaly recognition and adjustment schemes.

[0050] The risk assessment module is based on an anomaly identification and adjustment scheme. It uses a decision tree classifier to assess the loss risk during loading and transportation, and introduces risk factors into the cargo hold decision to generate risk assessment results.

[0051] The risk management strategy module, based on the risk assessment results, uses a Bayesian network to construct a probability model, comprehensively assesses the impact of cargo arrival delays and peak periods of uncertainty, provides a risk management basis for scheduling decisions, and generates risk factor management solutions.

[0052] The real-time optimization feedback module is based on a risk factor management scheme. It analyzes cabin allocation efficiency through real-time data monitoring and random forest algorithm, dynamically adjusts cabin allocation strategy, and optimizes it through simulated annealing algorithm to form closed-loop control and generate a real-time optimized cabin allocation strategy.

[0053] Compared with existing technologies, the beneficial effects of this invention are as follows: By applying principal component analysis and data cleaning methods, the accuracy and adaptability of handling complex relationships between cargo and cargo space are improved, enabling the cargo scheduling system to respond more flexibly to changes in market and operational conditions. Through refined management of cargo characteristics and cargo space conditions, more efficient cargo-cargo matching is achieved, thereby maximizing resource utilization efficiency. Contract design algorithms and mechanism design principles, along with dynamic programming algorithms and time-sensitive weighting methods, are employed to assign appropriate weights to cargoes with different time sensitivities and dynamically adjust their priorities. This not only improves the rapid response capability for urgent cargo but also ensures overall loading efficiency and reasonable scheduling of other cargoes, achieving a balance between efficiency and fairness. By analyzing historical cargo scheduling data and current scheduling plans through cluster analysis and the isolated forest algorithm, abnormal patterns can be identified and handled, improving the timeliness and effectiveness of handling abnormal situations. This also reduces the risk of resource waste or transportation delays caused by abnormal situations. Furthermore, the introduction of decision tree models and probabilistic models constructed using Bayesian networks effectively improves the ability to predict and manage cargo loss risks during loading and transportation. This not only reduces cargo loss but also optimizes loading plans. Finally, the application of real-time data monitoring methods and random forest algorithms improves the comprehensive management capabilities in the face of uncertainty. By collecting key performance indicators in real time and dynamically adjusting scheduling strategies, the system can respond more flexibly and accurately to uncertainties such as cargo arrival delays and sudden peak periods, ensuring the flexibility and accuracy of scheduling decisions. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart of the present invention;

[0056] Figure 2 This is a schematic diagram of step S1 in the intelligent cabin arrangement method of the present invention;

[0057] Figure 3 This is a schematic diagram of step S2 in the intelligent cabin arrangement method of the present invention;

[0058] Figure 4 This is a schematic diagram of step S3 in the intelligent cabin arrangement method of the present invention;

[0059] Figure 5 This is a schematic diagram of step S4 in the intelligent cabin arrangement method of the present invention;

[0060] Figure 6This is a schematic diagram of step S5 in the intelligent cabin arrangement method of the present invention;

[0061] Figure 7 This is a schematic diagram of step S6 in the intelligent cabin arrangement method of the present invention;

[0062] Figure 8 This is a schematic diagram of step S7 in the intelligent cabin arrangement method of the present invention;

[0063] Figure 9 This is a block diagram of the intelligent cabin sorting system of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Please see Figure 1 This invention provides a technical solution: an intelligent cabin sorting method, comprising the following steps:

[0066] S1: Based on the basic information of cargo and cargo space, principal component analysis and data cleaning methods are used to extract features and preprocess data, analyze cargo characteristics and demand, as well as cargo space conditions and constraints, and generate cargo space feature dataset.

[0067] S2: Based on the cargo space feature dataset, a cargo-space contract model is designed using contract design algorithms and mechanism design principles. The efficiency and fairness of the model are evaluated, and an optimized contract model set is generated.

[0068] S3: Based on the optimized contract model set, dynamic programming algorithm and time-sensitive weighted method are used to allocate weights to the cargo, dynamically adjust the cargo scheduling priority, and generate a time-priority scheduling plan.

[0069] S4: Based on the time-priority cabin allocation scheme, cluster analysis and isolated forest algorithms are used to analyze historical cabin allocation data and the current cabin allocation scheme, identify abnormal patterns, optimize the cabin allocation process, and generate anomaly identification and adjustment schemes.

[0070] S5: Based on the anomaly identification and adjustment scheme, a decision tree model is used to assess the risk of cargo loss during loading and transportation, and risk factors are introduced into the cargo allocation decision to generate risk assessment results.

[0071] S6: Based on the risk assessment results, a probability model is constructed using a Bayesian network. Taking into account the uncertainties of cargo arrival delays and sudden peak periods, a basis is provided for cargo scheduling decisions, and a risk factor management plan is generated.

[0072] S7: Based on the risk factor management scheme, the system uses real-time data monitoring and random forest algorithms to collect key performance indicators during the cabin allocation process in real time, dynamically adjusts the cabin allocation strategy, forms closed-loop control, and generates a real-time optimized cabin allocation strategy.

[0073] The cargo space feature dataset includes adjusted cargo volume parameters, weight parameters, sensitivity levels, space environmental adaptability parameters, and space configuration. The optimized contract model set includes cost-effectiveness ratios, contract default risk assessments, and time efficiency indicators customized for each cargo space combination. The time-priority scheduling scheme includes cargo priority ranking based on time sensitivity weighting, estimated arrival time, and emergency response level. The anomaly identification and adjustment scheme includes classification of abnormal scheduling events, preventive measures, and timeliness adjustment plans. The risk assessment results include cargo loss probability assessment, potential delay risk index, and safety measures. The risk factor management scheme includes uncertainty response strategies, risk mitigation measures, and emergency preparedness levels. The real-time optimized scheduling strategy includes adjusted loading order, optimized space allocation schemes, and real-time monitoring and feedback mechanisms.

[0074] In step S1, basic information about cargo and cargo space is processed using principal component analysis (PCA) and data cleaning methods. Specifically, data cleaning techniques are employed to assess data quality, and null value checking and duplicate record removal methods (such as the `isnull` and `drop_duplicates` functions) are used to clean the dataset, ensuring data accuracy and integrity. Z-score standardization is used to identify and handle outliers, guaranteeing data quality. Principal component analysis is performed, selecting principal components to reduce data dimensionality while retaining 95% of data variability. Specific PCA parameters, such as the percentage of variability to be retained (e.g., PCA(n_components=0.95)), are used to simplify the feature set and reduce subsequent computational complexity. Finally, a refined cargo space feature dataset is generated, containing adjusted cargo volume, weight parameters, and other key information, providing accurate input data for subsequent steps.

[0075] In step S2, a cargo-space contract model is designed based on the cargo space feature dataset using contract design algorithms and mechanism design principles. This process involves defining an initial contract design model, simulating basic terms such as transportation costs and time constraints using linear regression, and generating an initial contract model framework. Utilizing mechanism design principles, a game-theoretic optimizer is used to adjust model parameters (e.g., efficiency weight of 0.5 and fairness weight of 0.5). The optimized contract model undergoes a comprehensive evaluation of efficiency and fairness, ensuring optimal matching under information asymmetry. Through precise model design and evaluation, an optimized contract model set is generated, providing an efficient and fair solution for cargo and space matching.

[0076] In step S3, based on the optimized contract model set, dynamic programming and a time-sensitive weighted method are used to assign weights to goods and dynamically adjust their scheduling priorities. The dynamic programming environment is initialized, defining the state space and action space, corresponding to the scheduling priority and space allocation decisions for each type of goods, respectively. By assigning time-sensitive weights, reflecting the urgency of the goods, the dynamic programming algorithm dynamically adjusts the scheduling priority of goods according to these weights. This ensures that time-sensitive goods are processed first, effectively improving scheduling efficiency and response speed.

[0077] In step S4, historical cabin scheduling data and the current scheduling plan are analyzed using cluster analysis and the Isolation Forest algorithm to identify abnormal patterns. Cluster analysis groups historical cabin scheduling data to identify key scheduling patterns, while the Isolation Forest algorithm is used to detect abnormal cabin scheduling events that differ significantly from historical patterns. This optimizes the cabin scheduling process, improving its accuracy and efficiency by identifying and adjusting abnormal cabin scheduling events.

[0078] In step S5, a decision tree model is used to assess the risk of cargo loss during loading and transportation based on anomaly identification and adjustment schemes, and risk factors are incorporated into the scheduling decision. Data features associated with loss risk, such as cargo type and loading conditions, are collected, and the decision tree model is trained to predict loss risk. The resulting risk assessment results provide crucial risk information for scheduling decisions, helping to reduce cargo loss during transportation.

[0079] In step S6, a probabilistic model is constructed using a Bayesian network to comprehensively analyze uncertainties such as cargo arrival delays and sudden peak periods, providing a scientific basis for cargo scheduling decisions. The nodes and edges of the Bayesian network are defined, and the risks under different scenarios are calculated using conditional probability distributions to generate risk factor management plans. This improves the flexibility and accuracy of cargo scheduling decisions and effectively addresses uncertainties.

[0080] In step S7, key performance indicators (KPIs) during the cabin allocation process are collected using real-time data monitoring and a random forest algorithm, and the allocation strategy is dynamically adjusted. The real-time monitoring dataset includes indicators such as cabin utilization and loading speed. The random forest algorithm analyzes the data to identify abnormal efficiency patterns. This closed-loop control process generates a real-time optimized allocation strategy by dynamically adjusting the strategy, thereby improving the adaptability and efficiency of the allocation process.

[0081] Please see Figure 2 Based on the basic information of cargo and cargo space, principal component analysis and data cleaning methods are used to extract features and preprocess data, analyze cargo characteristics and demand, as well as cargo space conditions and limitations, and generate a cargo space feature dataset. The specific steps are as follows:

[0082] S101: Based on the basic information of cargo and cargo space, conduct data quality assessment, use the isnull() and drop_duplicates() methods to check and remove missing values ​​and duplicate records, apply the Z-score normalization method to identify and handle outliers, and generate a clean dataset;

[0083] S102: Based on a clean dataset, perform principal component analysis, use the PCA(n_components=0.95) command to select principal components, retain 95% of the data variability, reduce data dimensionality and retain key variables, and generate a feature set with simplified dimensions;

[0084] S103: Based on the simplified feature set, perform feature extraction, filter key parameters, including cargo volume, weight, type, as well as cargo bay size, location, and environmental conditions, to generate a cargo bay feature dataset.

[0085] In substep S101, data quality assessment and cleaning are performed by applying data cleaning techniques. This involves using the `isnull()` method to check for missing values ​​in the dataset and using the `drop_duplicates()` method to remove any duplicate records, ensuring the uniqueness and integrity of the data. Next, Z-score normalization is used to standardize the numerical features in the dataset. Outliers are identified and handled by calculating the difference between each value and the mean and dividing by the standard deviation. These operations ensure that the resulting dataset is clean, removing noise and inconsistencies, and providing accurate foundational data for subsequent analysis and processing.

[0086] In substep S102, principal component analysis (PCA) is performed to reduce the dimensionality of the dataset while preserving key information. The PCA(n_components=0.95) command is used, whose core objective is to select a sufficient number of principal components to retain 95% of the variability of the original data. By transforming high-dimensional data to a low-dimensional space while retaining as much information as possible, the feature set is effectively simplified. The dimensionality-simplified feature set generated in this step highlights the most important variables, laying the foundation for more efficient data processing and analysis.

[0087] In sub-step S103, feature extraction is performed to refine the selection of key parameters that significantly impact cargo allocation decisions. These parameters include cargo volume, weight, and type, as well as cargo space dimensions, location, and environmental conditions. By analyzing the simplified feature set, features that directly affect the cargo allocation process are identified and selected, forming a refined cargo space feature dataset. This dataset provides precise input for subsequent cargo-cargo space matching and contract model design, ensuring that the allocation system can formulate the most suitable allocation plan based on the actual characteristics of the cargo and the conditions of the cargo space, significantly improving allocation efficiency and accuracy.

[0088] Please see Figure 3 Based on a cargo space feature dataset, a cargo-space contract model is designed using contract design algorithms and mechanism design principles. The efficiency and fairness of the model are evaluated, and the specific steps for generating an optimized contract model set are as follows:

[0089] S201: Based on the cargo space feature dataset, define the initial model for contract design, use the LinearRegression() method to simulate basic terms, including transportation costs and time constraints, and generate an initial contract model framework.

[0090] S202: Based on the initial contract model framework, the GameTheoryOptimizer() method is used to optimize the model by applying the mechanism design principle and adjusting the parameters including efficiency_weight=0.5 and fairness_weight=0.5 to generate the optimized contract model.

[0091] S203: Based on the optimized contract model, a comprehensive evaluation is conducted. The evaluate_model(performance_metrics=[efficiency, fairness]) method is used to select the cargo-space contract model set and generate an optimized contract model set.

[0092] In sub-step S201, the initial contract design model is defined based on a cargo space feature dataset and uses linear regression to simulate transportation costs and time constraints. This involves taking features such as cargo volume, weight, and type, along with parameters such as space size, location, and environmental conditions, as inputs, and using a linear regression model to predict transportation costs and required time. The linear regression model provides a statistically based prediction for each cargo-space combination by analyzing the linear relationship between features and target variables (transportation costs and time). By accurately calculating transportation costs and time, an initial contract model framework is generated, providing a foundation for further contract optimization.

[0093] In substep S202, the initialized contract model is optimized using mechanism design principles and a game theory optimizer. Parameters, including efficiency weight and fairness weight, are adjusted and both set to 0.5 to balance efficiency and fairness in the contract model. The game theory optimizer simulates the interaction between cargo owners and space owners to find the contract model that maximizes overall utility. The contract conditions are optimized to ensure that the contract promotes efficient resource utilization while maintaining a balance of interests among the participating parties, thus generating an optimized set of contract models.

[0094] In sub-step S203, a comprehensive evaluation is performed based on the optimized contract model, using the evaluation model method (evaluate_model) to quantitatively assess the efficiency and fairness of the model. Performance indicators, including efficiency and fairness, are introduced, and specific evaluation methods are used to comprehensively evaluate the performance of each contract model. By comparing the evaluation results of different contract models, a set of contract models is selected that ensures both timely and efficient transportation of goods and fair treatment of the interests of all parties involved in the transportation process. The final optimized contract model set provides a scientific and reasonable basis for subsequent cargo allocation decisions, effectively improving the overall performance of the cargo allocation system.

[0095] Please see Figure 4 Based on the optimized contract model set, the dynamic programming algorithm and time-sensitive weighted method are used to allocate weights to the cargo, dynamically adjust the cargo scheduling priority, and generate a time-priority scheduling scheme. The specific steps are as follows:

[0096] S301: Based on the optimized contract model set, initialize the dynamic programming environment, set the state space as the priority of each type of cargo for storage, and the action space as the storage allocation decision. Use the initialize_dp_environment(states, actions) function to initialize and generate the dynamic programming initialization environment.

[0097] S302: Initialize the environment based on dynamic programming, apply the time-sensitive weighting method to assign weights to each type of goods, and use the assign_weights(data, time_sensitivity) function, where data is the characteristics of the goods and the time_sensitivity parameter is set according to the urgency of the goods to generate a time-sensitive weight allocation table.

[0098] S303: Based on the time-sensitive weight allocation table, execute the dynamic programming algorithm and dynamically adjust the cargo scheduling priority through the dynamic_programming(environment, weights) function, giving priority to time-sensitive cargo and generating a time-priority scheduling plan.

[0099] In substep S301, the initialization of the dynamic programming environment prepares the basic framework for subsequent operations. This includes using the output of the optimized contract model set—the detailed characteristics of each cargo and space combination—as input data. Using the `initialize_dp_environment` function, the state space is defined as the allocation priority for each cargo, while the action space represents the space allocation decision. This means the system considers the priority of each cargo and the space allocation scheme, generating an initial environment for each scenario. This provides a clear starting point for the dynamic programming algorithm, ensuring that subsequent steps can be performed within a precisely defined parameter space, thus laying the foundation for developing an effective allocation strategy.

[0100] In sub-step S302, a time-sensitivity weighted method is used to assign weights to each type of cargo, based on the cargo's urgency. The `assign_weights` function takes cargo characteristics (such as volume, weight, and type) and a time-sensitivity parameter as input, the latter set according to the cargo's urgency. This generates a time-sensitivity weight allocation table by assigning different weights to each type of cargo based on its time sensitivity. This ensures that the cargo scheduling system can identify and prioritize time-sensitive cargo, thereby improving efficiency while meeting customers' demands for rapid transportation.

[0101] In sub-step S303, a dynamic programming algorithm is executed to dynamically adjust the cargo scheduling priority based on a time-sensitivity weight allocation table. Using the `dynamic_programming` function, the algorithm calculates the optimal scheduling priority for each type of cargo using the dynamic programming initialization environment and time-sensitivity weight allocation table generated in the previous two sub-steps. Simultaneously, the algorithm considers the characteristics, time sensitivity, and available cargo space of various cargoes, ensuring that time-sensitive cargo receives priority processing. The resulting time-priority scheduling scheme optimizes the cargo scheduling order, ensuring that urgent cargo can be loaded and transported in the shortest possible time, significantly improving the efficiency of the overall scheduling process and customer satisfaction.

[0102] Please see Figure 5 Based on the time-first cabin allocation scheme, cluster analysis and the isolated forest algorithm are used to analyze historical cabin allocation data and the current cabin allocation scheme, identify abnormal patterns, optimize the cabin allocation process, and generate anomaly identification and adjustment schemes. The specific steps are as follows:

[0103] S401: Based on the time-priority cabin allocation scheme, cluster analysis is used to group historical cabin allocation data. The KMeans(n_clusters=3) function is used to select the number of clusters according to the data characteristics, identify three key cabin allocation patterns, and generate cabin allocation pattern clustering results.

[0104] S402: Based on the clustering results of the cabin allocation pattern, the IsolationForest (n_estimators=100, contamination=auto) algorithm is applied to detect anomalies in the current cabin allocation scheme, identify abnormal cabin allocation events that differ from historical patterns, and generate basic anomaly identification results.

[0105] S403: Based on the basic anomaly identification results, optimize and adjust the cabin scheduling process. Use the optimize_scheduling(report, constraints) method for optimization, formulate adjustment strategies, and generate anomaly identification and adjustment plans.

[0106] In sub-step S401, historical cargo scheduling data is grouped using cluster analysis. The KMeans clustering algorithm is employed, with the `n_clusters` parameter set to 3. This means that based on the characteristics of the historical scheduling data, the data is divided into three distinct clusters, identifying three key scheduling patterns. This method analyzes the scheduling time, priority, and space allocation characteristics of various goods, grouping similar scheduling situations into the same group to identify common scheduling patterns. This not only helps in understanding the patterns present in historical scheduling data but also provides a reference for current and future scheduling strategies. Through the generated scheduling pattern clustering results, scheduling decisions can more accurately match goods with space, improving scheduling efficiency.

[0107] In sub-step S402, based on the cabin allocation pattern clustering results, the Isolation Forest algorithm is used to detect anomalies in the current cabin allocation scheme. The Isolation Forest algorithm, with the n_estimators parameter set to 100 and the contamination parameter set to auto, aims to identify outliers in the data by constructing multiple isolated trees. The Isolation Forest algorithm identifies anomalies by isolating observations, making it suitable for quickly detecting unique cabin allocation events in large-scale data. The generated basic anomaly identification results enable the timely discovery of problems in the current cabin allocation scheme that significantly differs from historical allocation patterns, providing crucial information for optimizing the cabin allocation process.

[0108] In sub-step S403, the scheduling process is optimized based on the basic anomaly identification results. The `optimize_scheduling` method is employed, which utilizes data from the anomaly identification results, combined with reports and constraints in the scheduling process, to optimize the scheduling strategy. This optimization process considers various constraints, such as cabin capacity, loading limits, and time windows, aiming to develop an adjustment plan that not only addresses identified anomalies but also improves overall scheduling efficiency and customer satisfaction. The generated anomaly identification and adjustment plan not only solves the problems in the current scheduling scheme but also optimizes the entire scheduling process, ensuring smoother and more efficient scheduling operations.

[0109] Please see Figure 6 Based on the anomaly identification and adjustment scheme, a decision tree model is used to assess the risk of cargo loss during loading and transportation, and risk factors are introduced into the cargo allocation decision. The specific steps to generate risk assessment results are as follows:

[0110] S501: Based on the anomaly identification and adjustment scheme, collect and organize data features associated with loss risk, including organizing cargo type, loading conditions and historical loss records from the anomaly identification and adjustment scheme, creating a data frame using the pandas.DataFrame method, and generating a risk assessment dataset;

[0111] S502: Based on the risk assessment dataset, a decision tree model is used for training. The model is trained using DecisionTreeClassifier(criterion=gini, max_depth=4) to generate a loss risk assessment model.

[0112] S503: Based on the loss risk assessment model, the model.predict(features) method is used to assess the loss risk of each cargo in the current cargo allocation plan and generate the risk assessment results.

[0113] In the S501 submodule, a risk assessment dataset is constructed by collecting data features associated with loss risk. This includes key information such as cargo type, loading conditions, and historical loss records, gathered from anomaly identification and adjustment plans. Using the DataFrame method of the pandas library, the data is formatted into a structured data frame for easy subsequent processing. This not only reflects the risk factors faced by goods during transportation but also provides an empirical basis for the risk assessment model, ensuring the accuracy and reliability of the assessment results.

[0114] In the S502 submodule, a decision tree model is used to train the risk assessment dataset, aiming to establish a model capable of predicting cargo loss risk. Using DecisionTreeClassifier, the Gini coefficient is selected as the splitting criterion, and the maximum depth is set to 4 to avoid overfitting while ensuring sufficient learning depth. This training process generates a decision tree model that can assess loss risk based on cargo characteristics and loading conditions by analyzing the relationship between features in the dataset and loss risk. This model serves as the core tool for loss risk assessment, providing a scientific risk evaluation basis for cargo allocation decisions.

[0115] In the S503 submodule, based on a pre-trained loss risk assessment model, the `model.predict` method is used to assess the loss risk of each cargo in the current cargo scheduling plan. This involves inputting the cargo's feature data into a decision tree model, which then predicts the loss risk level of each cargo based on previously trained rules. The generated risk assessment results detail the loss risks faced by each type of cargo, providing crucial reference information for scheduling decisions. This allows the scheduling process to take risk factors into account more precisely, optimize cargo loading and transportation plans, reduce potential losses, and improve the efficiency and safety of the entire logistics process.

[0116] Please see Figure 7Based on the risk assessment results, a probability model is constructed using a Bayesian network. Taking into account the uncertainties of cargo arrival delays and sudden peak periods, this model provides a basis for scheduling decisions. The specific steps for generating a risk factor management plan are as follows:

[0117] S601: Based on the risk assessment results, define the nodes and edges of the Bayesian network, build the model using the BayesianModel method of the pgmpy library, the nodes include the risk of delayed arrival of goods, the risk of sudden peak period, and the risk of loss of goods, and set the dependencies between the nodes to generate the Bayesian network structure.

[0118] S602: Based on the Bayesian network structure, the model is trained by a Bayesian estimator. The fit() method is used, with the input parameter being the dataset of risk assessment results. The conditional probability distribution between nodes is calculated, and the Bayesian network after parameter training is generated.

[0119] S603: Based on the Bayesian network after parameter learning, the probabilistic inference query() method is executed to evaluate the impact of cargo arrival delays and sudden peak periods on loss risk and generate risk factor management plans.

[0120] In substep S601, a probabilistic model is constructed by defining the nodes and edges of a Bayesian network, including using the BayesianModel method from the pgmpy library. The data format is primarily based on risk assessment results, including cargo arrival delay risk, sudden peak period risk, and cargo loss risk as network nodes. By defining the dependencies between nodes—that is, defining which nodes directly affect the probabilities of other nodes—a complete Bayesian network structure is established. This reflects the interactions between different risk factors, providing a powerful tool for in-depth analysis. Constructing such a network allows for more accurate identification and analysis of key risk factors influencing cargo scheduling decisions, providing a foundation for developing effective risk management strategies.

[0121] In substep S602, a Bayesian estimator is used to learn the parameters of the Bayesian network, primarily through the `fit` method. The input parameters are the risk assessment dataset collected and organized in the previous step, which contains actual observations of different risk factors. By calculating the conditional probability distributions between nodes, the parameter-learned Bayesian network can more accurately reflect the relationships between various risk factors in reality. The generated Bayesian network model not only theoretically establishes the dependencies between risk factors but also possesses parameters learned from actual data, improving the accuracy and reliability of risk prediction.

[0122] In substep S603, probabilistic inference is performed, employing a query method to assess the impact of cargo arrival delays and sudden peak periods on loss risk. Based on a Bayesian network with learned parameters, the probability distribution of unknown risk factors is calculated using known risk factor values. This probabilistic inference can predict the changing trend of cargo loss risk under specific conditions, providing a scientific risk assessment for scheduling decisions. The generated risk factor management plan details the scheduling strategies to be adopted at different risk levels, helping decision-makers make more rational scheduling decisions when facing uncertainties, effectively reducing cargo loss risk and ensuring the safety and efficiency of the transportation process.

[0123] Please see Figure 8 Based on the risk factor management scheme, the real-time data monitoring method and random forest algorithm are used to collect key performance indicators in the cabin allocation process in real time, dynamically adjust the cabin allocation strategy, form closed-loop control, and generate a real-time optimized cabin allocation strategy. The specific steps are as follows:

[0124] S701: Based on the risk factor management scheme, real-time data collection is carried out. The sensor_data=collect_real_time_data(sensors) function is used to collect data on cabin utilization and loading speed, and generate a real-time monitoring dataset.

[0125] S702: Based on the real-time monitoring dataset, the RandomForestClassifier(n_estimators=100, max_features=auto, random_state=42) algorithm is used to analyze the data, identify the abnormal cabin departure patterns, and generate abnormal efficiency pattern identification results.

[0126] S703: Based on the results of efficiency anomaly pattern recognition, dynamically adjust the cabin scheduling strategy, use the adjust_scheduling(strategy, feedback) function to adjust the strategy, apply the simulated annealing algorithm to optimize the adjusted cabin scheduling strategy, and generate a real-time optimized cabin scheduling strategy.

[0127] In sub-step S701, a real-time data collection mechanism is implemented to focus on acquiring key performance indicators during the cabin clearing process, such as cabin utilization and loading speed. Data is collected from various sensors using a sensor data collection function (sensor_data=collect_real_time_data) and formatted into a real-time monitoring dataset. This dataset records the real-time changes in cabin utilization and loading speed in time-series format, providing fundamental data for identifying abnormal patterns in cabin clearing efficiency. The collected real-time data directly reflects the efficiency of the current cabin clearing operation, allowing adjustments to the clearing strategy to be based on the latest operational data, thereby improving the responsiveness and adaptability of the overall cabin clearing process.

[0128] In substep S702, the Random Forest algorithm is used to analyze the real-time monitoring dataset, aiming to identify anomaly patterns in efficiency during the cabin depot process. The Random Forest algorithm, by setting 100 decision trees (n_estimators=100) and allowing the model to automatically adjust when selecting the optimal features (max_features=auto), ensures the model's generalization ability and accuracy. The setting of a random state (random_state=42) guarantees the repeatability of the results. This enables the accurate identification of anomalous events that significantly differ from historical cabin depot efficiency from a large amount of real-time data. The generated efficiency anomaly pattern identification results provide timely feedback to the cabin depot process, ensuring the efficiency and reliability of the cabin depot operation.

[0129] In sub-step S703, the scheduling strategy is dynamically adjusted based on the efficiency anomaly pattern identification results. The `adjust_scheduling` function, combined with a simulated annealing algorithm, is used to optimize the scheduling strategy. The simulated annealing algorithm finds the global optimum by simulating the annealing process in the physical process, avoiding the problem of traditional optimization algorithms easily getting trapped in local optima. This strategy adjustment is not only based on the current scheduling efficiency and the identified anomaly patterns, but also considers the overall goals of the scheduling operation, such as improving bay utilization and shortening loading time. By dynamically optimizing the scheduling strategy in this way, the generated real-time optimized scheduling strategy aims to improve scheduling efficiency, reduce the impact of abnormal events, and ensure that the scheduling operation can quickly adapt to various emergencies, achieving closed-loop control of the scheduling process.

[0130] Please see Figure 9 An intelligent cabin allocation system includes a data quality assessment module, a feature engineering and optimization module, a contract model design module, a cabin allocation priority adjustment module, an anomaly pattern recognition module, a risk assessment module, a risk management strategy module, and a real-time optimization feedback module.

[0131] The data quality assessment module uses basic information about cargo and cargo space to clean and preprocess data by employing methods such as null value checking and deduplication, as well as outlier handling techniques, to generate a clean dataset.

[0132] The feature engineering and optimization module, based on a clean dataset, reduces data dimensionality while retaining key variables through principal component analysis and feature selection techniques, extracts key features of cargo and cargo hold, and generates a cargo hold feature dataset.

[0133] The contract model design module is based on a cargo hold feature dataset. It uses linear regression to simulate basic terms and combines the Nash equilibrium solution method to balance efficiency and fairness, thereby optimizing the contract model and generating an optimized contract model set.

[0134] The cargo priority adjustment module is based on an optimized contract model set. It applies dynamic programming and time-sensitive weighted methods to assign priorities to cargo, dynamically adjusts the cargo allocation order, and generates a time-priority cargo allocation scheme.

[0135] The anomaly pattern recognition module is based on the time-first cabin scheduling scheme. It uses cluster analysis and the isolated forest algorithm to identify anomaly patterns in the cabin scheduling data, optimizes the cabin scheduling process, and generates anomaly recognition and adjustment schemes.

[0136] The risk assessment module is based on an anomaly identification and adjustment scheme. It uses a decision tree classifier to assess the loss risk during loading and transportation, and introduces risk factors into the cargo hold decision to generate risk assessment results.

[0137] The risk management strategy module, based on the risk assessment results, uses a Bayesian network to build a probability model, comprehensively evaluates the impact of cargo arrival delays and peak periods of uncertainty, provides a risk management basis for scheduling decisions, and generates risk factor management plans.

[0138] The real-time optimization feedback module is based on the risk factor management scheme. It analyzes cabin allocation efficiency through real-time data monitoring and random forest algorithm, dynamically adjusts cabin allocation strategy, and optimizes it through simulated annealing algorithm to form closed-loop control and generate real-time optimized cabin allocation strategy.

[0139] The data quality assessment module ensures the accuracy and integrity of data, providing a reliable foundation for all subsequent decisions. Clean datasets reduce the possibility of erroneous decisions and improve the overall decision-making quality of the system. The feature engineering and optimization module, through refined principal component analysis and feature selection, ensures that only key variables are used in the decision-making process, significantly reducing computational complexity while preserving decision accuracy. This not only improves the system's operational efficiency but also ensures the quality of cargo allocation plans, optimizing resource utilization. The innovative application of the contract model design module, such as combining linear regression and Nash equilibrium methods, balances efficiency and fairness, providing a scientific basis for matching cargo with cargo space. This optimized design not only improves the efficiency of resource allocation but also ensures that the interests of all parties are reasonably considered, enhancing the system's acceptability and fairness. The cargo allocation priority adjustment module's dynamic programming and time-sensitive weighted methods provide a flexible adjustment mechanism for cargo loading order, ensuring that urgent cargo can be prioritized, improving the responsiveness to customer needs and the timeliness of service. The introduction of the anomaly pattern recognition module enables the system to promptly detect and adjust abnormal scheduling events, avoiding potential losses and risks, and enhancing the stability and reliability of the scheduling process. The combined use of the risk assessment and risk management strategy modules provides comprehensive risk considerations and management strategies for scheduling decisions, reducing risks during transportation, ensuring cargo safety, and improving customer satisfaction. Finally, the real-time optimization feedback module, through a closed-loop control mechanism, achieves real-time monitoring and dynamic adjustment of the scheduling strategy, ensuring that the scheduling strategy is always in an optimal state and improving the adaptability and efficiency of the entire scheduling system.

[0140] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0141] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart cabin arrangement method, characterized in that, Includes the following steps: Based on the basic information of cargo and cargo space, principal component analysis and data cleaning methods are used to extract features and preprocess data, analyze cargo characteristics and demand, as well as cargo space conditions and constraints, and generate a cargo space feature dataset. Based on the aforementioned cargo space feature dataset, a cargo-space contract model is designed using contract design algorithms and mechanism design principles. The efficiency and fairness of the model are evaluated, and an optimized contract model set is generated. Based on the optimized contract model set, a dynamic programming algorithm and a time-sensitive weighted method are used to allocate weights to the cargo, dynamically adjust the cargo scheduling priority, and generate a time-priority scheduling scheme. Based on the time-priority cabin allocation scheme, cluster analysis and the isolated forest algorithm are used to analyze historical cabin allocation data and the current cabin allocation scheme, identify abnormal patterns, optimize the cabin allocation process, and generate anomaly identification and adjustment schemes. Based on the aforementioned anomaly identification and adjustment scheme, a decision tree model is used to assess the risk of cargo loss during loading and transportation, and risk factors are introduced into the cargo allocation decision to generate risk assessment results. Based on the risk assessment results, a probability model is constructed using a Bayesian network. Taking into account the uncertainties of cargo arrival delays and sudden peak periods, a basis is provided for cargo scheduling decisions, and a risk factor management plan is generated. Based on the aforementioned risk factor management scheme, a real-time data monitoring method and a random forest algorithm are used to collect key performance indicators in the cabin allocation process in real time, dynamically adjust the cabin allocation strategy, form a closed-loop control, and generate a real-time optimized cabin allocation strategy. Using contract design algorithms and mechanism design principles, a cargo-space contract model is designed, and the efficiency and fairness of the model are evaluated. The specific steps to generate an optimized contract model set are as follows: Based on the cargo space feature dataset, an initial model for contract design is defined, and the LinearRegression() method is used to simulate basic terms, including transportation costs and time constraints, to generate an initial contract model framework. Based on the initialized contract model framework, applying the mechanism design principle, the GameTheoryOptimizer() method is used to optimize the model, adjusting parameters including efficiency_weight=0.5 and fairness_weight=0.5, to generate the optimized contract model; Based on the optimized contract model, a comprehensive evaluation is conducted. The evaluate_model(performance_metrics=[efficiency, fairness]) method is used to select the cargo-space contract model set and generate an optimized contract model set. The specific steps for generating a time-priority cargo scheduling scheme are as follows: Using dynamic programming algorithm and time-sensitive weighted method, cargo is weighted and dynamically adjusted to adjust its scheduling priority. Based on the optimized contract model set, the dynamic programming environment is initialized, the state space is set as the priority of each type of cargo for booking, and the action space is set as the booking decision. The initialize_dp_environment(states, actions) function is used for initialization to generate the dynamic programming initialization environment. Based on the dynamic programming initialization environment, a time-sensitive weighting method is applied to assign weights to each type of goods using the assign_weights(data, time_sensitivity) function, where data represents the characteristics of the goods and the time_sensitivity parameter is set according to the urgency of the goods, to generate a time-sensitive weight allocation table. Based on the time-sensitive weight allocation table, a dynamic programming algorithm is executed. The dynamic_programming(environment, weights) function is used to dynamically adjust the cargo scheduling priority, giving priority to time-sensitive cargo, and generating a time-priority scheduling plan.

2. The intelligent cabin arrangement method according to claim 1, characterized in that: The cargo space feature dataset includes adjusted cargo volume parameters, weight parameters, sensitivity levels, space environmental adaptability parameters, and space configuration. The optimized contract model set includes cost-effectiveness ratios, contract default risk assessments, and time efficiency indicators customized for each cargo space combination. The time-priority scheduling scheme includes cargo priority ranking based on time sensitivity weighting, estimated arrival time, and emergency response level. The anomaly identification and adjustment scheme includes classification of abnormal scheduling events, preventive measures, and timeliness adjustment plans. The risk assessment results include cargo loss probability assessment, potential delay risk index, and safety measures. The risk factor management scheme includes uncertainty response strategies, risk mitigation measures, and emergency preparedness levels. The real-time optimized scheduling strategy includes adjusted loading order, optimized space allocation schemes, and real-time monitoring and feedback mechanisms.

3. The intelligent cabin arrangement method according to claim 1, characterized in that: Based on the basic information of cargo and cargo space, the specific steps for generating a cargo space feature dataset are as follows: using principal component analysis and data cleaning methods, feature extraction and data preprocessing are performed, cargo characteristics and demand are analyzed, as well as cargo space conditions and limitations are analyzed. Based on the basic information of cargo and cargo space, data quality is assessed. The isnull() and drop_duplicates() methods are used to check and remove missing values ​​and duplicate records. The Z-score normalization method is applied to identify and handle outliers, generating a clean dataset. Based on the clean dataset, principal component analysis was performed. The PCA(n_components=0.95) command was used to select principal components, retaining 95% of the data variability, reducing data dimensionality and retaining key variables, and generating a feature set with simplified dimensionality. Based on the simplified feature set, feature extraction is performed, and key parameters are selected, including the volume, weight, and type of cargo, as well as the size, location, and environmental conditions of the cargo hold, to generate a cargo hold feature dataset.

4. The intelligent cabin arrangement method according to claim 1, characterized in that: Based on the time-priority cabin allocation scheme, cluster analysis and the isolated forest algorithm are used to analyze historical cabin allocation data and the current cabin allocation scheme, identify abnormal patterns, optimize the cabin allocation process, and generate anomaly identification and adjustment schemes. The specific steps are as follows: Based on the time-priority cabin allocation scheme, cluster analysis is used to group historical cabin allocation data. The KMeans(n_clusters=3) function is used to select the number of clusters according to the data characteristics, identify three key cabin allocation patterns, and generate cabin allocation pattern clustering results. Based on the clustering results of the cabin allocation pattern, the IsolationForest (n_estimators=100, contamination=auto) algorithm is applied to detect anomalies in the current cabin allocation scheme, identify abnormal cabin allocation events that differ from historical patterns, and generate basic anomaly identification results. Based on the aforementioned basic anomaly identification results, the cabin scheduling process is optimized and adjusted using the optimize_scheduling(report, constraints) method. An adjustment strategy is then developed, and anomaly identification and adjustment schemes are generated.

5. The intelligent cabin arrangement method according to claim 1, characterized in that: Based on the aforementioned anomaly identification and adjustment scheme, the specific steps for using a decision tree model to assess the risk of cargo loss during loading and transportation, and introducing risk factors into the cargo allocation decision to generate risk assessment results are as follows: Based on the anomaly identification and adjustment scheme, data features associated with loss risk are collected and organized, including cargo type, loading conditions and historical loss records from the anomaly identification and adjustment scheme, and a data frame is created using the pandas.DataFrame method to generate a risk assessment dataset; Based on the aforementioned risk assessment dataset, a decision tree model is trained using DecisionTreeClassifier(criterion=gini, max_depth=4) to generate a loss risk assessment model. Based on the aforementioned loss risk assessment model, the model.predict(features) method is used to assess the loss risk of each cargo in the current cargo allocation plan and generate risk assessment results.

6. The intelligent cabin arrangement method according to claim 1, characterized in that: Based on the risk assessment results, a probability model is constructed using a Bayesian network. Taking into account the uncertainties of cargo arrival delays and sudden peak periods, a basis is provided for scheduling decisions. The specific steps for generating a risk factor management plan are as follows: Based on the risk assessment results, the nodes and edges of the Bayesian network are defined, and the BayesianModel method of the pgmpy library is used to build the model. The nodes include the risk of delayed arrival of goods, the risk of sudden peak period, and the risk of loss of goods. The dependencies between the nodes are set to generate the Bayesian network structure. Based on the Bayesian network structure, the model is trained by a Bayesian estimator. The fit() method is used, with the input parameter being the dataset of the risk assessment results. The conditional probability distribution between nodes is calculated to generate the Bayesian network after parameter training. Based on the Bayesian network learned from the parameters, the probabilistic inference query() method is executed to evaluate the impact of cargo arrival delays and sudden peak periods on loss risk, and generate a risk factor management plan.

7. The intelligent cabin arrangement method according to claim 1, characterized in that: Based on the aforementioned risk factor management scheme, the specific steps for using real-time data monitoring methods and random forest algorithms to collect key performance indicators during the cabin allocation process in real time, dynamically adjust the cabin allocation strategy, form closed-loop control, and generate a real-time optimized cabin allocation strategy are as follows: Based on the aforementioned risk factor management scheme, real-time data collection is performed. The sensor_data=collect_real_time_data(sensors) function is used to collect data on cabin utilization and loading speed, generating a real-time monitoring dataset. Based on the real-time monitoring dataset, the RandomForestClassifier(n_estimators=100, max_features=auto, random_state=42) algorithm is used to analyze the data, identify the abnormal cabin arrangement patterns, and generate abnormal efficiency pattern identification results. Based on the efficiency anomaly pattern recognition results, the cabin scheduling strategy is dynamically adjusted. The adjust_scheduling(strategy, feedback) function is used to adjust the strategy, and the simulated annealing algorithm is applied to optimize the adjusted cabin scheduling strategy to generate a real-time optimized cabin scheduling strategy.

8. An intelligent cabin sorting system, characterized in that: The intelligent cabin allocation method according to any one of claims 1-7 is implemented, wherein the system includes a data quality assessment module, a feature engineering and optimization module, a contract model design module, a cabin allocation priority adjustment module, an anomaly pattern recognition module, a risk assessment module, a risk management strategy module, and a real-time optimization feedback module; The data quality assessment module uses basic information about cargo and cargo space to clean and preprocess data by employing methods such as null value checking and deduplication, as well as outlier handling techniques, to generate a clean dataset. The feature engineering and optimization module is based on a clean dataset. Through principal component analysis and feature selection techniques, it reduces data dimensionality while retaining key variables, extracts key features of cargo and cargo space, and generates a cargo space feature dataset. The contract model design module is based on a cargo hold feature dataset. It uses linear regression to simulate basic terms and combines the Nash equilibrium solution method to balance efficiency and fairness, thereby optimizing the contract model and generating an optimized contract model set. The cargo allocation priority adjustment module is based on an optimized contract model set. It applies dynamic programming and time-sensitive weighted methods to allocate priorities to cargo, dynamically adjusts the cargo allocation order, and generates a time-priority cargo allocation scheme. The abnormal pattern recognition module is based on the time-first cabin scheduling scheme. It uses cluster analysis and the isolated forest algorithm to identify abnormal patterns in the cabin scheduling data, optimizes the cabin scheduling process, and generates anomaly recognition and adjustment schemes. The risk assessment module is based on an anomaly identification and adjustment scheme. It uses a decision tree classifier to assess the loss risk during loading and transportation, and introduces risk factors into the cargo hold decision to generate risk assessment results. The risk management strategy module, based on the risk assessment results, uses a Bayesian network to construct a probability model, comprehensively assesses the impact of cargo arrival delays and peak periods of uncertainty, provides a risk management basis for scheduling decisions, and generates risk factor management solutions. The real-time optimization feedback module is based on a risk factor management scheme. It analyzes cabin allocation efficiency through real-time data monitoring and random forest algorithm, dynamically adjusts cabin allocation strategy, and optimizes it through simulated annealing algorithm to form closed-loop control and generate a real-time optimized cabin allocation strategy.

Images