Logistics network reconstruction method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of logistics network reconstruction, and discloses a logistics network reconstruction method and device, electronic equipment and a storage medium, and the method comprises the steps: constructing a space-time thermodynamic diagram based on related geopolitical events; constructing a port layer penetration model, an airspace layer penetration model and a land layer penetration model, performing cascade connection to generate a combined model, and converting the combined model into a mixed integer linear programming model; and solving through a preset cascade analysis solving algorithm to obtain a reconstruction scheme. The method has the beneficial effects that the space-time thermodynamic diagram and the penetration model are constructed, interconnection among layers can quickly evaluate and adjust a transportation scheme when an event occurs, the stability and the efficiency of a supply chain are ensured, the feasibility and the efficiency of a reconstruction scheme are improved by utilizing the optimization capability of mixed integer linear programming, and the method is suitable for large-scale popularization and application. Scientific and data-driven strategy selection is provided for decision makers, and adaptability and anti-pressure ability of the global logistics network are enhanced.

Description

Technical Field

[0001] This invention relates to the field of logistics network reconstruction technology, and in particular to a logistics network reconstruction method, apparatus, electronic device and storage medium. Background Technology

[0002] Traditional logistics management models typically rely on historical data and fixed statistical analyses, failing to reflect the immediate impact of sudden geopolitical events. This inability to cope with unforeseen circumstances leads to the collapse of logistics networks and waste of resources. For example, when conflict suddenly erupts in a region, relevant transport routes may be blocked, and the uncertainty of alternative routes and the real-time requirements make traditional models inefficient. This not only affects the efficiency of cargo transportation but also increases labor costs and operational risks. Furthermore, when dealing with complex international relations, traditional methods fail to fully consider unstructured information such as political risks and market sentiment, resulting in biased and delayed decision-making. Summary of the Invention

[0003] Based on this, it is necessary to propose a logistics network reconstruction method, device, electronic equipment and storage medium to address the existing logistics network reconstruction problem.

[0004] A method for reconstructing a logistics network, the method comprising: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

[0005] Furthermore, the step of constructing a spatiotemporal heatmap based on the relevant geopolitical events if such events exist includes: Key elements of the relevant geopolitical events were extracted using a named entity recognition model. Based on the aforementioned key elements, multiple target dimension data are obtained to form a target dimension dataset. The target dimension dataset is scored by a preset multidimensional evaluation model to obtain the heat values ​​corresponding to the key elements; The spatiotemporal heat map is constructed based on the key elements and the thermal values.

[0006] Furthermore, before the step of scoring the target dimension dataset using a preset multidimensional evaluation model to obtain the heatmap values ​​corresponding to the key elements, the method further includes: Obtain the correlation values ​​between the relevant geopolitical events and each target dimension in the preset multidimensional assessment model; The scoring weights of each target dimension of the multidimensional evaluation model are set based on the correlation values.

[0007] Furthermore, the step of solving the mixed-integer linear programming model using a preset cascaded analysis algorithm to obtain the reconstruction scheme includes: Output the Pareto solution set based on the aforementioned mixed-integer linear programming model; Chromosome encoding is performed on the schemes in the Pareto solution set to obtain the first chromosome set; Perform a crossover mutation operation on the first chromosome set to obtain a second chromosome set; The second chromosome set is combined with the first chromosome set to generate a third chromosome set; Using the third chromosome set as the first chromosome set, the target step and the steps following the target step are repeated until the number of chromosomes in the third chromosome set is greater than or equal to a preset number; the target step is to perform a crossover mutation operation on the first chromosome set to obtain the second chromosome set. The third chromosome set whose number of chromosomes is greater than or equal to the preset number is denoted as the target third chromosome set; The fitness value of each chromosome in the target third chromosome set is calculated using a preset fitness calculation function; The chromosome with the highest fitness value was selected as the reconstruction scheme.

[0008] Furthermore, before the step of outputting the Pareto solution set based on the mixed-integer linear programming model, the method further includes: Obtain the constraints; The constraints are input into the mixed-integer linear programming model to obtain a first objective mixed-integer linear programming model; wherein the first objective mixed-integer linear programming model is used to output the Pareto solution set.

[0009] Furthermore, before the step of outputting the Pareto solution set based on the mixed-integer linear programming model, the method further includes: Define the objective function; The objective function is input into the mixed-integer linear programming model to obtain a second-objective mixed-integer linear programming model; wherein the second-objective mixed-integer linear programming model is used to output the Pareto solution set.

[0010] Furthermore, prior to the steps of real-time detection of relevant geopolitical events in the current logistics network and real-time acquisition of datasets from three preset dimensions (sea, land, and air), the following steps are included: Raw data from various databases is crawled using a distributed crawler cluster. A multilingual bill of lading parsing engine built using the XLM-RoBERTa model is used to parse the original data to obtain a geopolitical event set; wherein, the geopolitical event set is used to detect whether the relevant geopolitical events exist.

[0011] A logistics network restructuring device, the device comprising: The detection module is used to detect in real time whether there are relevant geopolitical events in the current logistics network, and to acquire datasets in three preset dimensions: sea, land, and air. The first construction module is used to construct a spatiotemporal heat map based on the relevant geopolitical events if such events exist. The second construction module is used to construct port layer penetration model, airspace layer penetration model and land route layer penetration model respectively based on the datasets of each preset dimension and the spatiotemporal heat map; The cascading module is used to cascade the port layer penetration model, the airspace layer penetration model, and the land route layer penetration model to generate a combined model. The conversion module is used to convert the combined model into a mixed integer linear programming model; The calculation module is used to solve the mixed integer linear programming model through a preset cascade analysis solution algorithm to obtain a reconstruction scheme.

[0012] An electronic device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

[0013] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

[0014] The beneficial effects of this invention are as follows: by constructing a spatiotemporal heat map and a penetration model, the interconnection between different levels can quickly assess and adjust transportation plans when events occur, ensuring the stability and efficiency of the supply chain. By utilizing the optimization capabilities of mixed-integer linear programming, the feasibility and efficiency of the reconfiguration plan are improved, providing decision-makers with scientific and data-driven strategy choices and enhancing the adaptability and resilience of the global logistics network. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0016] in: Figure 1 This is a diagram illustrating the application environment of a logistics network reconstruction method in one embodiment. Figure 2 Here is a flowchart of a logistics network reconstruction method in one embodiment; Figure 3 This is a structural block diagram of a logistics network reconfiguration device in one embodiment; Figure 4This is a structural block diagram of an electronic device in one embodiment. Detailed Implementation

[0017] 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.

[0018] Figure 1 This is a diagram illustrating the application environment of logistics network reconstruction in one embodiment. (Refer to...) Figure 1 This logistics network reconstruction method is applied to a logistics network reconstruction system. The system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal; a mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of multiple servers. The terminal 110 is used to acquire relevant geopolitical events, and the server 120 is used to generate a logistics network reconstruction scheme.

[0019] like Figure 2 As shown, in one embodiment, a logistics network reconstruction method is provided. This method can be applied to both terminals and servers; this embodiment illustrates its application to terminals. The logistics network reconstruction method specifically includes the following steps: S1: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; S2: If the relevant geopolitical events are present, then construct a spatiotemporal heat map based on the relevant geopolitical events; S3: Construct port layer penetration model, airspace layer penetration model and land route layer penetration model based on the datasets of each preset dimension and the spatiotemporal heat map respectively; S4: Cascade the port layer penetration model, the airspace layer penetration model, and the land route layer penetration model to generate a combined model; S5: Transform the combined model into a mixed integer linear programming model; S6: Solve the mixed integer linear programming model using a preset cascade analysis solution algorithm to obtain a reconstruction scheme.

[0020] As described in step S1 above, the system needs to detect in real time whether there are relevant geopolitical events in the current logistics network, and to acquire datasets from three preset dimensions: sea, land, and air. Specifically, a multi-source data access system needs to be built to identify and confirm the occurrence of geopolitical events through various information channels (such as social media monitoring, news analysis, government reports, etc.), including monitoring events such as conflicts, sanctions, and changes in trade policies. Simultaneously, the system also needs to continuously acquire datasets from the three preset dimensions related to logistics: sea, land, and air. The sea-based dataset includes shipping data (such as ship AIS signals), ship waiting times, and crane operating efficiency. The land-based dataset includes border crossing efficiency, average truck dwell time at border crossings, and real-time GPS data. The air-based dataset includes air traffic data, ADS-B signals and NOTAMs (Notifications to Airmen), and no-fly zone extension paths. In addition, each dataset can also include relevant climate, economic, and political environmental indicators.

[0021] As described in step S2 above, if relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. The purpose of the spatiotemporal heat map is to visualize the impact of geopolitical events on the logistics network and its evolution. Based on the acquired data, the intensity of logistics activities, pressure indicators, and risk levels in each region within a specific time period are calculated. Furthermore, algorithms such as kernel density estimation can be used to transform this data into a visual heat map, representing the pressure, risk, and activity intensity of different regions with different colors and graphics. The visual heat map can not only help decision-makers quickly identify high-risk areas but also display potential logistics bottlenecks and key nodes.

[0022] As described in step S3 above, port layer penetration models, airspace layer penetration models, and land layer penetration models are constructed based on the datasets of each preset dimension and the spatiotemporal heatmap. The port layer penetration model focuses on evaluating the port's throughput capacity and its carrying capacity under the influence of geopolitical events. The model needs to consider factors such as port congestion and the availability of alternative ports. The airspace layer penetration model mainly analyzes situations where air transport is restricted, such as the impact of no-fly zones or route changes on air transport efficiency, and identifies potential alternative routes. The land layer penetration model mainly focuses on the availability of ground transport channels, assessing border crossing capacity and the risks of land transport. Each model should integrate its specific data source and real-time information obtained through spatiotemporal heatmap analysis to ensure the model's timeliness and accuracy. Specifically, after obtaining their respective data sources, the port layer penetration model can be constructed using any of the following modeling software: AnyLogic (discrete event simulation), FlexSim (3D port modeling), and PortOpt (dedicated port optimization tool). The data sources for the port layer penetration model can include the following data: infrastructure dimension: number of berths / maximum draft / quay crane efficiency; operation dimension: TEU turnover rate (container volume per berth per hour); geopolitical risk coefficient: political stability index × port strategic value weight. The airspace layer penetration model can be constructed using any of the following modeling software: AirTop (air traffic control simulation) and SkyBrea. The (fuel optimization), Python+PostGIS (custom development), the data source for the airspace layer penetration model can include the following data: spatial dimension: 3D airway corridor (including altitude layer), time dimension: peak traffic period mapping, dynamic heat map overlay: using QGIS spatiotemporal cube to display the no-fly zone diffusion pattern; the land layer penetration model can be constructed using any of the following modeling software: TransCAD (dedicated to transportation planning), Vissim (microscopic simulation), Arena (supply chain risk assessment), the data source for the land layer penetration model can include the following data: daily number of China-Europe freight trains, alternative routes, geopolitics, natural disasters, social unrest.

[0023] As described in step S4 above, the port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model. Specifically, this is based on the interrelationships and dependencies between different layers. For example, port congestion may affect airspace and land route logistics arrangements, and changes in airspace may affect land route transportation choices. Therefore, the cascaded model must ensure information flow between the three layers so that changes in the state of each layer can affect other layers. By establishing multi-layered interactive relationships, the system can more comprehensively assess the impact of geopolitical events on the entire logistics network, thereby forming a holistic, real-time response mechanism. Specifically, the specific steps for generating the combined model are as follows: 1) Determine the model parameters and variables; Determine the basic variables: Transportation mode selection variables: For example, define Choose the transportation option (sea, air, rail, etc.) from node i to node j.

[0024] Resource allocation variables: such as This represents the amount of warehousing resources located in a specific geographical location.

[0025] Constraints: Define the constraints for each layer, such as port throughput, number of flights in the airspace, and land traffic capacity.

[0026] 2) Establish the output relationships of each layer of the model; Port layer penetration model output: Congestion rate, throughput, and availability of alternative ports at the port of export.

[0027] Output of the spatial layer penetration model: The feasibility of routes with limited capacity and alternative routes, and the potential impact on the number of flights and time delays.

[0028] For example, if an area is marked as a no-fly zone, calculate the additional transit time and cost required.

[0029] Land layer penetration model output: Assessment of the capacity of land transport, the volume of goods at the border, and potential obstacles.

[0030] For example, based on historical data, assess the saturation time of a border crossing's transshipment capacity.

[0031] 3) Integrate interactive relationships; Define the effects of interaction: Clarify the interactions between different layers, such as whether port congestion affects flight adjustments, or whether airspace restrictions alter land transport decisions.

[0032] For example, the output of the port layer can be incorporated into the model construction of the airspace layer and the land route layer using a matrix form.

[0033] 4) Establish a feedback mechanism: Define the feedback relationships between different model layers, such as how to influence airspace and land route optimization schemes in real time when congestion occurs in the port layer.

[0034] For example, if it is anticipated that a port will reach its maximum capacity, the airspace and land transport layers should immediately take appropriate measures to change transport strategies.

[0035] 5) Construct an overall composite model; Combinatorial model design: The results of the three-layer model are integrated into a unified model framework, which can include the decision variables and constraints of each layer, forming a complete logical network that allows information to flow between layers.

[0036] 6) Verify and adjust the model; Model validation: The model is validated using historical data and simulated events to ensure that the combined model accurately reflects real-world situations and can handle various possibilities.

[0037] Compare the model's predicted results with the actual results to check for significant differences, and adjust the parameters or algorithm as needed.

[0038] Model Adjustment: Make necessary adjustments based on the validation results, such as resetting some constraints or optimizing the model output based on feedback information.

[0039] Ensure that the entire combined model remains adaptable and robust in the context of geopolitical events.

[0040] As described in step S5 above, the combined model is transformed into a mixed-integer linear programming model. Mixed-integer linear programming is an optimization method applicable to decision-making in resource allocation and logistics network planning. This stage requires clarifying each decision variable, such as transportation mode selection, resource allocation, and cost control, and establishing corresponding linear constraints. These constraints need to be set based on the output of the penetration model and actual operational constraints (such as transportation capacity and time windows). By formatting the complex cascading analysis into a mathematical model, the system can provide decision-makers with clear optimization objectives, making the reconstructed scheme feasible.

[0041] As described in step S6 above, the mixed-integer linear programming model is solved using a preset cascaded analysis algorithm to obtain a reconstruction scheme. The mixed-integer linear programming model is solved using a preset cascaded analysis algorithm, and a reconstruction scheme for the logistics network is obtained in the process. This process uses optimization algorithms (such as branch and bound algorithm, simplex method, etc.) to find the optimal solution of the model. The solution results will show the optimal resource allocation and transportation choices under different scenarios, and specific logistics adjustment measures can be formulated based on these results. This provides support for subsequent implementation, enabling the logistics network to be quickly adjusted under the influence of geopolitical events, ensuring the continuity and efficiency of its operation.

[0042] In one embodiment, step S2, which involves constructing a spatiotemporal heatmap based on the relevant geopolitical events if such events exist, includes: S201: Extract key elements of the relevant geopolitical events using a named entity recognition model; S202: Based on the key elements, obtain multiple target dimension data to obtain a target dimension dataset; S203: The target dimension dataset is scored using a preset multidimensional evaluation model to obtain the heat values ​​corresponding to the key elements; S204: Construct the spatiotemporal heat map based on the key elements and the thermal values.

[0043] As described in steps S201-S204 above, key elements of relevant geopolitical events are extracted using a Named Entity Recognition (NER) model. Named Entity Recognition is a natural language processing technique designed to identify specific units of information, such as people, locations, organizations, and significant events, from text. Various information sources, such as news reports, social media activity, and policy statements, are monitored in real time, and the NER model is used to extract feature information related to geopolitical events. For example, when a country announces sanctions, the model can extract key elements such as the country's name, affected countries, type of sanctions, and implementation time. These extracted key elements provide the foundation for subsequent data acquisition and analysis, helping to establish background knowledge of the current event and accurately define the scope of impact, thus laying the foundation for the model's real-time performance and accuracy. Based on these elements, data from multiple target dimensions are acquired to form a target dimension dataset, which includes multiple dimensions such as economic indicators, transportation conditions, social security indices, historical conflict records, and weather data. Through cross-analysis of multi-source data, the system can reflect the potential impact of geopolitical events on these dimensions. For example, if a conflict causes logistics delays in a certain region, the system needs to collect information such as traffic flow data, port throughput, and flight cancellations in that region.

[0044] After acquiring the target dimension dataset, a pre-defined multidimensional assessment model is used to score the data, deriving heatmap values ​​related to geopolitical events. The multidimensional assessment model can be any of the following algorithms, such as weighted average, analytic hierarchy process (AHP), or fuzzy comprehensive evaluation. These algorithms calculate a comprehensive score based on the weights assigned to each dimension of the data. The heatmap value is a comprehensive indicator, typically used to represent the degree of risk or impact intensity faced by a specific region within a given timeframe. In this way, the system can quantify the contribution of each dimension to the overall risk. For example, if the social security index of a region significantly declines, its heatmap value will be correspondingly increased. By combining key elements with heatmap values, the system constructs a corresponding spatiotemporal heatmap. A spatiotemporal heatmap is a visualization tool that can show the distribution of data in time and space. In some embodiments, key elements can also be mapped onto a Geographic Information System (GIS), assigning different color depths to different regions based on the intensity of the heatmap values, with darker areas representing areas with higher impact. Rendering techniques and graphics processing algorithms can be used to enhance the visualization of the heatmap, allowing users to easily identify the risk level of key areas. Meanwhile, such visualization also helps decision-makers intuitively understand the impact of geopolitical events and adjust their response strategies in a timely manner. This process not only improves the operability of the model but also enhances the timeliness and accuracy of overall decision-making, providing intuitive and powerful support for subsequent logistics network restructuring.

[0045] In one embodiment, before step S203, which involves scoring the target dimension dataset using a preset multidimensional evaluation model to obtain the heat values ​​corresponding to key elements, the method further includes: S2021: Obtain the correlation values ​​between the relevant geopolitical events and each target dimension in the preset multidimensional assessment model; S2022: Set the scoring weights for each target dimension of the multidimensional evaluation model based on the correlation values.

[0046] As described in steps S2021-S2022 above, before scoring the heat map value, it is necessary to first analyze the correlation values ​​between relevant geopolitical events and the various target dimensions in the preset multidimensional assessment model. This determines how different dimensions are affected by specific geopolitical events; the correlation value can be calculated using any method, such as correlation coefficient analysis, regression analysis, or pattern recognition based on historical data. Through the analysis of historical events, the system identifies past situations with similar patterns to the current geopolitical event, thereby assigning corresponding correlation values ​​to the target dimensions. For example, analyzing a supply chain disruption event caused by sanctions in a certain country may reveal significant correlations between multiple dimensions such as economic indicators, traffic delays, and social stability and the event. Based on this, the system can generate a correlation value for each target dimension, representing its importance and degree of influence in the current event. The setting of this correlation value will directly affect the dynamic balance of subsequent scoring, ensuring that the model fully considers the close connection with actual events when evaluating the heat map value.

[0047] After obtaining the correlation values ​​of target dimensions related to geopolitical events, corresponding scoring weights are assigned to each target dimension of the multidimensional evaluation model based on these correlation values. To ensure fairness and scientific rigor, the system generally uses a normalization method to convert the correlation values ​​into weights, preventing any single dimension from dominating the overall score due to an excessively large numerical range. The process of setting weights may include setting the ratio of maximum to minimum values ​​to ensure the rationality of weight allocation, and methods such as weighted average and analytic hierarchy process (AHP) can be used for weight calculation. For example, the entropy weight method can be used to calculate the weights of each dimension. Let i be the information entropy of the i-th dimension, and n be the total number of dimensions. Through these methods, the system can better reflect the importance of each dimension in the context of the current geopolitical event. For example, if the correlation value of a certain dimension is much higher than that of other dimensions, then its corresponding weight will be higher, meaning that this dimension has a greater influence on the overall result when calculating the heatmap value. At the same time, the weight setting must also consider the possible independence and interaction between different dimensions to avoid potential scoring bias. Through this process, the final weight configuration enables subsequent heatmap value calculations to reflect the actual impact of each dimension in the specific event context, contributing to the formation of a more accurate and efficient evaluation model.

[0048] In one embodiment, step S6, which involves solving the mixed-integer linear programming model using a preset cascaded analysis algorithm to obtain a reconstruction scheme, includes: S601: Output the Pareto solution set based on the mixed-integer linear programming model; S602: Chromosome encoding is performed on the schemes in the Pareto solution set to obtain the first chromosome set; S603: Perform a crossover mutation operation on the first chromosome set to obtain a second chromosome set; S604: Combine the second chromosome set with the first chromosome set to generate a third chromosome set; S605: Using the third chromosome set as the first chromosome set, repeat the target step and the steps following the target step until the number of chromosomes in the third chromosome set is greater than or equal to a preset number; the target step is to perform a crossover mutation operation on the first chromosome set to obtain the second chromosome set. S606: The third chromosome set with a number of chromosomes greater than or equal to a preset number is denoted as the target third chromosome set; S607: Calculate the fitness value of each chromosome in the target third chromosome set using a preset fitness calculation function; S608: Select the chromosome with the highest fitness value as the reconstruction scheme.

[0049] As described in steps S601-S605 above, the system outputs a Pareto solution set based on the solution results of the Mixed Integer Linear Programming (MILP) model. In multi-objective optimization problems, the Pareto solution set represents the solution where no further improvement to any objective is possible without sacrificing other objectives. It requires calculating performance indicators for each possible solution, such as transportation cost, timeliness, and risk value. By comprehensively evaluating these indicators, the system can identify all optimal solutions that achieve a balance between different objectives, thus forming a Pareto front in a multi-dimensional space. In one embodiment, a heuristic algorithm can be used to pre-screen feasible solutions to reduce the scale of the MILP solution. Each solution corresponds to a practical logistics restructuring scheme with different characteristics and possible application scenarios. For example, the solution with the lowest transportation cost may not necessarily be the one with the best timeliness, but rather a reasonable balance between cost and timeliness. This process provides multiple excellent initial solutions for subsequent evolutionary algorithms. Through appropriate solution methods, the system can improve the quality of the solutions through multiple iterations.

[0050] Each solution in the Pareto solution set is encoded into a chromosome, forming the first chromosome set. The purpose of chromosome encoding in this step is to transform the optimized solution into a format suitable for genetic algorithm operations. Typically, chromosomes are represented as binary strings, arrays of real numbers, or other encoding methods. For example, the decision variables for each solution (such as the choice of transportation mode, the allocation of resources, etc.) can be transformed into a series of genes, forming a chromosome. For instance, the port selection, freight rate, and transportation plan for a certain solution can be encoded as "110101," where each bit represents a different choice. In this way, the system can effectively handle these complex solutions while facilitating subsequent crossover and mutation operations. Ensuring the quality and integrity of the encoding is crucial for the success of subsequent evolutionary operations; good encoding preserves the diversity of information and the possibility of mutation, achieving the optimization goal. Crossover and mutation operations are performed on the first chromosome set to generate the second chromosome set. Crossover and mutation are core operations in genetic algorithms, aiming to explore better solutions by simulating the natural selection process. In the crossover operation, two chromosomes are randomly selected and crossed at a certain position to generate new offspring chromosomes. This process helps combine information from the parent chromosomes to explore possible new solutions. For example, parent chromosomes "110101" and "011000" might generate "110000" and "011101" after crossover. Subsequent mutation operations randomly alter a gene on the chromosome to introduce new features or options, increasing the diversity of the search space. Through these operations, the generated second chromosome set not only maintains excellent parental features but also effectively increases new search paths, avoiding the risk of early convergence to local optima. It should be noted that the probability of mutation for each gene segment can be set, for example, to 15%. The second chromosome set is then combined with the first chromosome set to generate a third chromosome set. Through this merging operation, the system can retain information from the original high-quality solutions while introducing newly generated solutions, thereby enhancing the diversity of the solution space. This ensures that excellent genetic features are not lost in any iteration. The third chromosome set is used as the new first chromosome set, and then the target steps are repeated iteratively. The target steps mainly refer to performing crossover and mutation operations on the chromosome sets to obtain a new second chromosome set. This iterative process continues until the number of chromosomes in the resulting third chromosome set is greater than or equal to a preset number. Through multiple generations of evolution and genetic operations, the system can continuously explore new solutions and optimize them repeatedly to ensure the diversity and quality of the solutions. This process not only increases the breadth of knowledge but also enhances the effectiveness of the evolutionary algorithm. This iterative strategy allows the algorithm to perform a comprehensive search in the solution space, helping to find the optimal or near-optimal solution, while ensuring that enough solutions are generated for decision-makers to choose from in practical applications.

[0051] As described in steps S606-S608 above, a third chromosome set with a number of chromosomes greater than or equal to a preset number is denoted as the target third chromosome set. The fitness value of each chromosome in the target third chromosome set is calculated using a preset fitness calculation function. The fitness value is a quantitative indicator that measures the performance of each solution under the objective function, and is usually combined with multi-objective optimization problems, such as a comprehensive evaluation of factors like cost, time, and risk. The system will evaluate the performance of each chromosome solution in a given environment by analyzing resource allocation, transportation efficiency, and related costs based on the preset fitness function. This process may involve multi-objective trade-offs to ensure a balance between different indicators. For example, the fitness function may impose a higher weight on transportation costs to reflect the importance of cost control in the current market environment. The chromosome with the highest fitness value is selected as the reconstruction scheme. This achieves the final goal of multi-objective optimization, ensuring that the selected scheme achieves the best effect after comprehensive consideration of cost, timeliness, and risk. This selection process usually combines the previously set objective weights and fitness function to ensure that the selected scheme not only meets business needs but also has practicality and feasibility. The final restructuring plan will provide decision-makers with a clear action plan to ensure that logistics networks can respond quickly and effectively to changes under the influence of geopolitical events, thereby reducing related risks.

[0052] In one embodiment, before step S601 of outputting the Pareto solution set based on the mixed-integer linear programming model, the method further includes: S6001: Obtain constraints; S6002: Input the constraints into the mixed-integer linear programming model to obtain a first objective mixed-integer linear programming model; wherein the first objective mixed-integer linear programming model is used to output the Pareto solution set.

[0053] As described in steps S6001-S6002 above, before solving the mixed-integer linear programming (MILP) model, it is necessary to first collect and organize all constraints related to the optimization problem. Constraints are key factors defining the feasible solution set in the model, typically including resource limitations, operational conditions, time constraints, and legal regulations. For example, capacity constraints: the maximum carrying capacity of each transportation node (such as a port or warehouse) cannot exceed its actual capacity. Time constraints: transported goods must be completed within a specified time, with minimum time limits between delivery points. Budget constraints: the total project cost must be lower than or equal to a preset funding limit, including transportation costs and tariffs. Regulatory constraints: compliance with international and local trade laws and regulations, such as the use of logistics equipment and related documentation requirements. After obtaining the constraints, these constraints are input into the mixed-integer linear programming model to construct a new model, called the first-objective mixed-integer linear programming model. The core objective of this model is to solve a specific optimization problem and output the corresponding Pareto solution set. During the process of inputting constraints into the model, the system needs to convert each constraint into a mathematical expression to form an operable form.

[0054] In one embodiment, before step S601, which outputs the Pareto solution set based on the mixed-integer linear programming model, the method further includes: S6101: Define the objective function; S6102: Input the objective function into the mixed-integer linear programming model to obtain a second objective mixed-integer linear programming model; wherein the second objective mixed-integer linear programming model is used to output the Pareto solution set.

[0055] As described in steps S6101-S6102 above, a clear objective function is set before solving the mixed-integer linear programming (MILP) model. It should be noted that the MILP model can also pre-set its objective function; this can be done by updating the objective function. The objective function is a key component of the optimization model, defining the core indicators the model aims to optimize. In logistics and supply chain management, the objective function typically involves multiple aspects, such as: cost minimization: reducing overall transportation costs, including cargo transportation, warehousing, and distribution costs, by selecting the most economical transportation methods and routes. Service level maximization: improving customer service quality, such as ensuring on-time delivery rates or increasing customer satisfaction. Risk minimization: considering potential risk factors in decision-making to ensure reduced potential losses in response to uncertainties (such as geopolitical events, market fluctuations, etc.). When setting the objective function, multiple objectives usually need to be considered comprehensively, especially when multi-objective optimization is involved, involving the setting of objective weights to reasonably balance multiple objectives during the solution process. This objective function is input into the mixed-integer linear programming model to construct a second-objective mixed-integer linear programming model.

[0056] In one embodiment, before step S1, which involves real-time detection of relevant geopolitical events in the current logistics network and real-time acquisition of datasets across three preset dimensions (sea, land, and air), the following steps are included: S001: Crawling raw data from various databases using a distributed crawler cluster; S002: A multilingual bill of lading parsing engine built using the XLM-RoBERTa model parses the original data to obtain a geopolitical event set; wherein, the geopolitical event set is used to detect whether the relevant geopolitical events exist.

[0057] As described in steps S001-S002 above, before implementing real-time detection of geopolitical events, the first step is to crawl raw data from various databases using a distributed crawler cluster. This process is fundamental to data collection and involves automatically extracting information from various online data sources using web crawling technology, including news websites, social media, government announcements, and reports from international organizations. The distributed crawler cluster consists of multiple crawler nodes that can work in parallel, significantly improving the efficiency of data acquisition. It processes real-time data streams using streaming computing frameworks (such as Apache Storm) with a response latency of less than one minute. For example, one crawler node might be responsible for accessing a specific news website, while another node focuses on social media platforms. Through its distributed design, the system can efficiently handle a large number of requests, avoiding the limitations encountered from frequently requesting a single website. Simultaneously, crawling frequency and time limits can be set as needed to ensure the real-time nature and up-to-dateness of the data. When crawling data, it is important to ensure that the extracted data is structured or semi-structured for subsequent natural language processing (NLP) and analysis. Meanwhile, the quality of the raw data directly affects the results of subsequent analysis. Therefore, it is crucial to ensure that the collected data is reliable, relevant, and stored appropriately for future use. A multilingual bill of lading parsing engine, built using the XLM-RoBERTa model, parses the collected raw data to identify and extract geopolitical events. XLM-RoBERTa is a pre-trained multilingual transformer trained on large-scale multilingual datasets through unsupervised learning, effectively handling text in multiple languages. Using this model, the system can automatically identify key elements of geopolitical events in the text, such as conflict, sanctions, and negotiations. This may involve extracting specific keywords, entities (such as countries and organizations), and event descriptions. The parsing engine transforms the raw data into a structured set of geopolitical events, containing detailed information about each event, including its type, time, location, and participants. Through automated bill of lading parsing, the system can promptly capture global geopolitical dynamics, providing strong support for addressing uncertainty and optimizing decision-making.

[0058] Reference Figure 3 The present invention also provides a logistics network restructuring device, the device comprising: The detection module 902 is used to detect in real time whether there are relevant geopolitical events in the current logistics network, and to acquire datasets in three preset dimensions: sea, land and air. The first construction module 904 is used to construct a spatiotemporal heat map based on the relevant geopolitical events if such events exist. The second construction module 906 is used to construct port layer penetration model, airspace layer penetration model and land route layer penetration model respectively based on the datasets of each preset dimension and the spatiotemporal heat map; The cascading module 908 is used to cascade the port layer penetration model, the airspace layer penetration model, and the land route layer penetration model to generate a combined model. The conversion module 910 is used to convert the combined model into a mixed integer linear programming model; The calculation module 912 is used to solve the mixed integer linear programming model through a preset cascade analysis solution algorithm to obtain a reconstruction scheme.

[0059] In one embodiment, the first building module 904 includes: The key element extraction submodule is used to extract key elements of the relevant geopolitical events through a named entity recognition model. The target dimension data acquisition submodule is used to acquire multiple target dimension data based on the key elements to obtain a target dimension dataset. The scoring submodule is used to score the target dimension dataset using a preset multidimensional evaluation model to obtain the heat values ​​corresponding to the key elements. The spatiotemporal heat map construction submodule is used to construct the spatiotemporal heat map based on the key elements and the heat values.

[0060] In one embodiment, the first building module 904 further includes: The correlation value acquisition submodule is used to acquire the correlation values ​​between the relevant geopolitical events and each target dimension in the preset multidimensional assessment model; The scoring weight setting submodule is used to set the scoring weights for each target dimension of the multidimensional evaluation model based on the correlation value.

[0061] In one embodiment, the computing module 912 includes: The Pareto solution set output submodule is used to output a Pareto solution set based on the mixed-integer linear programming model. The chromosome encoding submodule is used to encode the schemes in the Pareto solution set into chromosomes, thereby obtaining the first chromosome set; The mutation submodule is used to perform a crossover mutation operation on the first chromosome set to obtain a second chromosome set; The combination submodule is used to combine the second chromosome set with the first chromosome set to generate a third chromosome set; An iterative submodule is used to repeatedly execute the target step and the steps following the target step, using the third chromosome set as the first chromosome set, until the number of chromosomes in the third chromosome set is greater than or equal to a preset number; the target step is to perform a crossover mutation operation on the first chromosome set to obtain the second chromosome set. The tagging submodule is used to mark the third chromosome set whose number of chromosomes is greater than or equal to a preset number as the target third chromosome set; The fitness value calculation submodule is used to calculate the fitness value of each chromosome in the target third chromosome set through a preset fitness calculation function; The selection submodule is used to select the chromosome with the highest fitness value as the reconstruction scheme.

[0062] In one embodiment, the computing module 912 further includes: The constraint acquisition submodule is used to acquire constraints. The constraint input submodule is used to input the constraint conditions into the mixed-integer linear programming model to obtain a first objective mixed-integer linear programming model; wherein the first objective mixed-integer linear programming model is used to output the Pareto solution set.

[0063] In one embodiment, the computing module 912 further includes: The objective function setting submodule is used to set the objective function; The objective function input submodule is used to input the objective function into the mixed-integer linear programming model to obtain a second objective mixed-integer linear programming model; wherein the second objective mixed-integer linear programming model is used to output the Pareto solution set.

[0064] In one embodiment, a logistics network reconfiguration apparatus includes: The raw data crawling module is used to crawl raw data from various databases through a distributed crawler cluster; The raw data parsing module is used to parse the raw data using a multilingual bill of lading parsing engine built with the XLM-RoBERTa model to obtain a geopolitical event set; wherein, the geopolitical event set is used to detect whether the relevant geopolitical events exist.

[0065] Figure 4 An internal structural diagram of an electronic device in one embodiment is shown. This electronic device can specifically be a terminal or a server, and more specifically, a computer device. Figure 4As shown, the electronic device includes a processor, a memory, and a network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement a logistics network reconstruction method. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to implement the logistics network reconstruction method. Those skilled in the art will understand that... Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0066] In one embodiment, an electronic device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

[0067] By constructing spatiotemporal heat maps and penetration models, the interconnections between different levels can quickly assess and adjust transportation plans when events occur, ensuring the stability and efficiency of the supply chain. By utilizing the optimization capabilities of mixed-integer linear programming, the feasibility and efficiency of reconfiguration plans are improved, providing decision-makers with scientific and data-driven strategy choices and enhancing the adaptability and resilience of the global logistics network.

[0068] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, causes the processor to perform the following steps: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

[0069] By constructing spatiotemporal heat maps and penetration models, the interconnections between different levels can quickly assess and adjust transportation plans when events occur, ensuring the stability and efficiency of the supply chain. By utilizing the optimization capabilities of mixed-integer linear programming, the feasibility and efficiency of reconfiguration plans are improved, providing decision-makers with scientific and data-driven strategy choices and enhancing the adaptability and resilience of the global logistics network.

[0070] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0071] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0072] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for reconstructing a logistics network, characterized in that, The method includes: Real-time detection of relevant geopolitical events in the current logistics network, and real-time acquisition of datasets from three preset dimensions: sea, land, and air; If the aforementioned relevant geopolitical events exist, a spatiotemporal heat map is constructed based on these events. Based on the datasets of each of the preset dimensions and the spatiotemporal heatmap, port layer penetration model, airspace layer penetration model and land route layer penetration model are constructed respectively. The port layer penetration model, the airspace layer penetration model, and the land route layer penetration model are cascaded to generate a combined model; The combined model is transformed into a mixed integer linear programming model; The mixed-integer linear programming model is solved by a preset cascaded analysis algorithm to obtain a reconstruction scheme.

2. The logistics network reconstruction method according to claim 1, characterized in that, The step of constructing a spatiotemporal heatmap based on the relevant geopolitical events, if such events exist, includes: Key elements of the relevant geopolitical events were extracted using a named entity recognition model. Based on the aforementioned key elements, multiple target dimension data are obtained to form a target dimension dataset. The target dimension dataset is scored by a preset multidimensional evaluation model to obtain the heat values ​​corresponding to the key elements; The spatiotemporal heat map is constructed based on the key elements and the thermal values.

3. The logistics network reconstruction method according to claim 2, characterized in that, Before the step of scoring the target dimension dataset using a preset multidimensional evaluation model to obtain the heat values ​​corresponding to the key elements, the method further includes: Obtain the correlation values ​​between the relevant geopolitical events and each target dimension in the preset multidimensional assessment model; The scoring weights of each target dimension of the multidimensional evaluation model are set based on the correlation values.

4. The logistics network reconstruction method according to claim 1, characterized in that, The step of solving the mixed-integer linear programming model using a preset cascaded analysis algorithm to obtain a reconstruction scheme includes: Output the Pareto solution set based on the aforementioned mixed-integer linear programming model; Chromosome encoding is performed on the schemes in the Pareto solution set to obtain the first chromosome set; Perform a crossover mutation operation on the first chromosome set to obtain a second chromosome set; The second chromosome set is combined with the first chromosome set to generate a third chromosome set; Using the third chromosome set as the first chromosome set, the target step and the steps following the target step are repeated until the number of chromosomes in the third chromosome set is greater than or equal to a preset number; the target step is to perform a crossover mutation operation on the first chromosome set to obtain the second chromosome set. The third chromosome set whose number of chromosomes is greater than or equal to the preset number is denoted as the target third chromosome set; The fitness value of each chromosome in the target third chromosome set is calculated using a preset fitness calculation function; The chromosome with the highest fitness value was selected as the reconstruction scheme.

5. The logistics network reconstruction method according to claim 4, characterized in that, Before the step of outputting the Pareto solution set based on the mixed-integer linear programming model, the method further includes: Obtain the constraints; The constraints are input into the mixed-integer linear programming model to obtain a first objective mixed-integer linear programming model; wherein the first objective mixed-integer linear programming model is used to output the Pareto solution set.

6. The logistics network reconstruction method according to claim 4, characterized in that, Before the step of outputting the Pareto solution set based on the mixed-integer linear programming model, the method further includes: Define the objective function; The objective function is input into the mixed-integer linear programming model to obtain a second-objective mixed-integer linear programming model; wherein the second-objective mixed-integer linear programming model is used to output the Pareto solution set.

7. The logistics network reconstruction method according to claim 1, characterized in that, Before the steps of real-time detection of relevant geopolitical events in the current logistics network and real-time acquisition of datasets from three preset dimensions (sea, land, and air), the following are included: Raw data from various databases is crawled using a distributed crawler cluster. A multilingual bill of lading parsing engine built using the XLM-RoBERTa model is used to parse the original data to obtain a geopolitical event set; wherein, the geopolitical event set is used to detect whether the relevant geopolitical events exist.

8. A logistics network reconfiguration device, characterized in that, The device includes: The detection module is used to detect in real time whether there are relevant geopolitical events in the current logistics network, and to acquire datasets in three preset dimensions: sea, land, and air. The first construction module is used to construct a spatiotemporal heat map based on the relevant geopolitical events if such events exist. The second construction module is used to construct port layer penetration model, airspace layer penetration model and land route layer penetration model respectively based on the datasets of each preset dimension and the spatiotemporal heat map; The cascading module is used to cascade the port layer penetration model, the airspace layer penetration model, and the land route layer penetration model to generate a combined model. The conversion module is used to convert the combined model into a mixed integer linear programming model; The calculation module is used to solve the mixed integer linear programming model through a preset cascade analysis solution algorithm to obtain a reconstruction scheme.

9. A computer-readable storage medium, characterized in that, The system contains a computer program that, when executed by a processor, causes the processor to perform the steps of the logistics network reconfiguration method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the logistics network reconfiguration method as described in any one of claims 1 to 7.

Images