Risk prediction management system and method for cross-border supply chain

Abstract

The invention discloses a risk prediction management system and method for a cross-border supply chain, and belongs to the technical field of risk management, and the system comprises a plan decomposition module which is used for obtaining a sequential net demand; the supplier retrieval module is used for obtaining qualified suppliers; the task distribution module is used for obtaining a sequential production task sequence; the production compensation module is used for obtaining compensation time sequence production redundancy; the transportation compensation module is used for obtaining a compensation time sequence transportation task; the compensation fusion module is used for generating a first cross-border supply chain toughness strategy; and the risk avoidance module is used for executing active risk avoidance so as to solve the technical problem that full-link active risk avoidance is difficult to realize due to insufficient supply chain toughness caused by multi-source uncertainty of production and transportation links in a cross-border supply chain, and the purposes of compensating production redundancy and transportation congestion through dynamic prediction and improving the safety of the cross-border supply chain are achieved. The active risk avoidance capability and the resource optimization efficiency of the cross-border supply chain are improved, and the delivery delay and the operation cost are reduced.

Description

Technical Field

[0001] This invention relates to the field of risk management technology, specifically to a risk prediction management system and method for cross-border supply chains. Background Technology

[0002] As the global industrial chain continues to deepen, enterprises' production and operation activities increasingly rely on complex and lengthy cross-border supply chain networks. While this globalized procurement and production model optimizes resource allocation and reduces costs, it also introduces a large number of uncontrollable risks due to its numerous links and geographically dispersed nature. Therefore, traditional supply chain management models face problems such as insufficient precision in production demand decomposition, weak dynamic adaptability of suppliers, and high uncertainty in cross-border transportation in cross-border scenarios, making it difficult for the supply chain to meet the complex and ever-changing market demands. Existing technologies generally adopt static demand forecasting and single-dimensional risk assessment mechanisms, which cannot achieve precise time-series mapping of master production schedules to multi-level material demands, nor can they coordinate and compensate for supplier capacity fluctuations and cross-border logistics congestion, resulting in a transmission and amplification effect of risks at each link of the supply chain. Summary of the Invention

[0003] This application provides a risk prediction management system and method for cross-border supply chains, aiming to solve the technical problem that the multi-source uncertainty in the production and transportation links of cross-border supply chains leads to insufficient supply chain resilience and makes it difficult to achieve proactive risk avoidance throughout the entire process. The goal is to achieve the technical effect of improving the proactive risk avoidance capability and resource optimization efficiency of cross-border supply chains by dynamically predicting and compensating for production redundancy and transportation congestion, thereby reducing delivery delays and operating costs.

[0004] In view of the above problems, this application provides a risk prediction management system and method for cross-border supply chains.

[0005] The first aspect disclosed in this application provides a risk prediction and management system for cross-border supply chains. This system includes: a plan decomposition module for downloading the master production plan at the order management terminal and decomposing the master production plan based on the bill of materials hierarchy to obtain multiple time-series net demand quantities for various production demand materials; a supplier retrieval module for using the first production demand material and the first time-series net demand quantity as joint retrieval conditions to traverse the cross-border supplier qualification database and obtain K qualified suppliers; a task allocation module for allocating production tasks based on the first time-series net demand quantity and K real-time capacity constraints of the K qualified suppliers to obtain K time-series production task sequences; and a production compensation module for adjusting the production compensation based on the material production attributes of the K qualified suppliers. The K time-series production task sequences are predicted for production redundancy buffering, and production process compensation is performed based on the prediction results to obtain K compensated time-series production redundancies; the transportation compensation module is used to predict cross-border transportation congestion for the K time-series production task sequences based on the material transportation attributes of the K qualified suppliers, and to perform transportation time-series compensation based on the prediction results to obtain K compensated time-series transportation tasks; the compensation fusion module is used to fuse the K compensated time-series production redundancies and the K compensated time-series transportation tasks as a first cross-border supply chain resilience strategy; the risk avoidance module is used to construct a set of cross-border supply chain resilience strategies covering all material needs of production by analogy based on multiple time-series net demand quantities of the various production demand materials, and to execute proactive risk avoidance in all aspects of the cross-border supply chain.

[0006] Another aspect of this application discloses a risk prediction and management method for cross-border supply chains. This method includes: downloading the master production plan at the order management terminal and decomposing the master production plan based on the bill of materials hierarchy to obtain multiple time-series net requirements for various production demand materials; using the first production demand material and the first time-series net requirement as joint search conditions, traversing the cross-border supplier qualification database to obtain K qualified suppliers; allocating production tasks based on the first time-series net requirement and the K real-time capacity constraints of the K qualified suppliers to obtain K time-series production task sequences; and performing the K time-series production task sequences based on the material production attributes of the K qualified suppliers. The production redundancy buffer is predicted for the sequenced production task sequence, and production process compensation is performed based on the prediction results to obtain K compensated time-series production redundancies; cross-border transportation congestion is predicted for the K time-series production task sequences based on the material transportation attributes of the K qualified suppliers, and transportation time-series compensation is performed based on the prediction results to obtain K compensated time-series transportation tasks; the K compensated time-series production redundancies and K compensated time-series transportation tasks are integrated as the first cross-border supply chain resilience strategy; by analogy, a set of cross-border supply chain resilience strategies covering all material needs of production is constructed based on multiple time-series net demand quantities of various production demand materials, and proactive risk avoidance is implemented in all aspects of the cross-border supply chain.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The system employs a plan decomposition module to download the master production plan from the order management terminal and decomposes the master production plan hierarchically based on the bill of materials, resulting in multiple time-series net requirements for various production demand materials. A supplier retrieval module uses the first production demand material and the first time-series net requirement as joint search criteria to traverse the cross-border supplier qualification database, obtaining K qualified suppliers. A task allocation module allocates production tasks based on the first time-series net requirement and the K real-time capacity constraints of the K qualified suppliers, resulting in K time-series production task sequences. A production compensation module adjusts the production redundancy of the K time-series production task sequences based on the material production attributes of the K qualified suppliers. The system employs a buffer prediction mechanism and compensates for production process delays based on the prediction results, resulting in K compensated time-series production redundancies. A transportation compensation module predicts cross-border transportation congestion for the K time-series production tasks based on the material transportation attributes of the K qualified suppliers, and compensates for transportation delays based on the prediction results, resulting in K compensated time-series transportation tasks. A compensation fusion module integrates the K compensated time-series production redundancies and K compensated time-series transportation tasks as a first cross-border supply chain resilience strategy. A risk avoidance module, by analogy, constructs a set of cross-border supply chain resilience strategies covering all material requirements for production based on multiple time-series net demand quantities of various production demand materials, and executes proactive risk avoidance across the entire cross-border supply chain. This addresses the technical problem of insufficient supply chain resilience caused by multi-source uncertainties in the production and transportation links of cross-border supply chains, making it difficult to achieve proactive risk avoidance across all links. It achieves the technical effect of improving the proactive risk avoidance capability and resource optimization efficiency of cross-border supply chains, and reducing delivery delays and operating costs by dynamically predicting and compensating for production redundancy and transportation congestion.

[0009] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0010] Figure 1 This application provides a schematic diagram of a risk prediction and management system for cross-border supply chains.

[0011] Figure 2 This application provides a flowchart illustrating a risk prediction and management method for cross-border supply chains.

[0012] Explanation of reference numerals in the attached diagram: 11. Plan decomposition module; 12. Supplier retrieval module; 13. Task allocation module; 14. Production compensation module; 15. Transportation compensation module; 16. Compensation integration module; 17. Risk avoidance module. Detailed Implementation

[0013] This application addresses the technical problem of insufficient supply chain resilience caused by multi-source uncertainties in the production and transportation links of cross-border supply chains, making it difficult to achieve proactive risk avoidance across all links. It provides a risk prediction management system and method for cross-border supply chains, thereby improving the proactive risk avoidance capability and resource optimization efficiency of cross-border supply chains, and reducing delivery delays and operating costs by dynamically predicting and compensating for production redundancy and transportation congestion.

[0014] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0015] Example 1, as Figure 1 As shown in the embodiment of this application, a risk prediction and management system for cross-border supply chains is provided. The system includes:

[0016] The planning decomposition module 11 is used to download the master production plan from the order management terminal and decompose the master production plan based on the bill of materials hierarchy to obtain multiple time-sequential net requirements for various production demand materials.

[0017] Specifically, in the planning decomposition module 11, the system first retrieves the master production schedule (MPS) already established by the enterprise from the order management terminal via online download. This MPS is typically set by the enterprise based on actual demand and includes the production quantity of the target product and the corresponding delivery time requirements. Then, the system calls the corresponding bill of materials (BOM) for the product and decomposes the target product layer by layer into the required sub-components and basic raw materials based on this BOM. Then, according to the nested relationships within the BOM, the system expands the material requirements at each level downwards, establishing the dependency relationship from the target product to the bottom-level raw materials. Afterwards, based on this dependency relationship, combined with the production quantities of different batches in the MPS and the corresponding demand time windows, the system reverse-engineers the demand quantity and supply sequence of each material at different points in time, thus obtaining the net demand quantity of each production material at different time periods. These time-sequential net demand quantities not only clearly reflect the specific demand scale and time distribution of various materials within the production cycle but also provide basic data support for subsequent supplier retrieval, capacity constraint analysis, and task allocation, ensuring that the entire cross-border supply chain can operate in an orderly manner according to the MPS.

[0018] Furthermore, the plan decomposition module 11 includes:

[0019] The master production schedule is decomposed to obtain the time-series production quantity and time-series demand window of the target production product; the bill of materials is loaded according to the product code of the target production product, and a material hierarchy dependency relationship is constructed based on the multi-level nested structure of the bill of materials; based on the material hierarchy dependency relationship, the material demand back offset calculation of the time-series production quantity and time-series demand window is performed to obtain multiple time-series net demand quantities of the various production demand materials.

[0020] In a preferred embodiment, after the planning decomposition module 11 obtains the master production schedule, it parses the master production schedule according to preset key names to obtain the planned production quantity and corresponding delivery nodes of the target product in each time period. By chronologically arranging the parsed data, the chronological production quantity and chronological demand time window of the target product are obtained. Subsequently, based on the product code of the target product, the corresponding bill of materials (BOM) is called and loaded. This BOM is usually a multi-level nested structure, which can decompose the target product layer by layer into multiple sub-components, components, and down to the raw materials at the bottom level. Based on this BOM, the system can construct the hierarchical dependency relationship between materials, that is, determine the quantity ratio and assembly order between each type of raw material or sub-component and the upper-level product. For example, target product A is composed of component B and component E, component B is composed of raw materials C and D, and component E is composed of raw materials F and G. Subsequently, based on this hierarchical dependency, the time-series production quantity and time-series demand window are passed down layer by layer in reverse. During this process, the system calculates the required quantity of sub-components level by layer according to the production quantity of the target product and the usage ratio in the hierarchical dependency, and further extends this calculation to the raw material layer. For example, if producing one target product A requires two components B and one component E, each component B requires three raw materials C and two raw materials D, and each component E requires two raw materials F and two raw materials G, then in the case of planning to produce 100 units of target product A, it will be calculated that 200 components B and 100 components E are needed, along with 600 raw materials C, 400 raw materials D, 400 raw materials F, and 400 raw materials G. Simultaneously, the system combines the standard production cycle and lead time of each level of material to perform reverse offset calculations on the demand time window, thereby determining the demand time of each material at different time nodes. After completing the reverse offset calculation of the demand quantity, the calculation results will be arranged according to the demand time to form multiple time-series net demand quantities covering various production demand materials, providing input data for subsequent supplier retrieval and task allocation.

[0021] Furthermore, the plan decomposition module 11 includes:

[0022] Extract the first production quantity and the first demand time window from the time-series production quantity and time-series demand time window mapping; calculate the material demand reverse offset of the first production quantity based on the material hierarchy dependency relationship to obtain the first material gross demand; aggregate the first material gross demand to obtain multiple first net demands of the various production demand materials; compensate the first demand time window in advance according to the standard production time to output the first production time window; bind the first production time window to the multiple first net demands to obtain the first time-series net demand data group; after obtaining multiple time-series net demand data groups corresponding to the time-series demand time windows by analogy, reorganize the multiple time-series net demand data groups based on the various production demand materials to obtain multiple time-series net demands of the various production demand materials.

[0023] In one feasible implementation, the system first extracts time-series nodes one by one from the parsed time-series production quantities and time-series demand windows, mapping the extracted time-series nodes and their corresponding production quantities to a first production quantity and a first demand window. Then, using the constructed material hierarchy dependency relationships, the system performs the same reverse offset calculation of material demand on the first production quantity to deduce the quantities of various sub-components and raw materials required to meet the first production quantity, thus obtaining the corresponding first gross material demand. After obtaining the gross demand for various materials, the system deducts existing inventory, materials in transit, and available safety stock from the first gross material demand based on the data recorded in the inventory management module, thereby obtaining multiple first net demands for various production demand materials. These net demands represent the additional quantities that suppliers need to provide, accurately reflecting the supply chain gaps. Subsequently, based on the advance compensation time set in the standard production process, the first demand time window is adjusted in advance. For example, if component B requires a 2-week processing cycle, the system will shift the original delivery time forward by 2 weeks, generating a new first production time window. This production time window ensures that parts can be put into production of the next-level components or complete machines on time. Then, the system binds the obtained first production time window with multiple first net demand quantities one by one, forming a complete first time-sequential net demand data group. This first time-sequential net demand data group not only describes the quantity demand of materials but also clarifies the specific supply timing requirements. Further, by analogy with the above process, the same processing is applied to all time-sequential production quantities and time-sequential demand time windows in the master production schedule, generating corresponding time-sequential net demand data groups in sequence. Finally, based on various production demand materials, the multiple time-sequential net demand data groups are integrated to obtain multiple time-sequential net demand quantities covering the entire material level. These time-sequential net demand quantities will serve as the data basis for subsequent supplier retrieval, capacity matching, and risk prediction, ensuring dynamic coordination of the cross-border supply chain in both time and quantity.

[0024] The supplier search module 12 is used to use the first production demand material and the first time-series net demand as joint search conditions to traverse the cross-border supplier qualification database and obtain K qualified suppliers.

[0025] Specifically, in the supplier search module 12, after calculating the first time-series net demand, the first time-series net demand and the corresponding first production demand material are used as joint search conditions to traverse the cross-border supplier qualification database. This cross-border supplier qualification database is a pre-built supplier information management database that records multi-dimensional information such as the material supply scope, qualification certification status, historical performance, and capacity limits of different suppliers. By comparing the joint search conditions with the corresponding data in the cross-border supplier qualification database, K suppliers that meet the joint search condition requirements can be selected and these suppliers are designated as qualified suppliers. This ensures the matching of materials in terms of quantity and delivery time, providing reliable input for subsequent production task allocation and risk compensation prediction, thereby effectively improving the robustness and flexibility of the cross-border supply chain.

[0026] Furthermore, the supplier search module 12 includes:

[0027] Extract the capacity matching constraint interval from the first time-series net demand; use the material code of the first production demand material and the first net demand interval as joint search conditions to traverse the cross-border supplier qualification database to obtain the K qualified suppliers; wherein, the qualification certification of the qualified suppliers includes the first production demand material, and the upper limit of capacity elasticity covers the capacity matching constraint interval.

[0028] In a preferred embodiment, the net demand for each material is first extracted from the first time-series net demand. Then, the net demand is multiplied by a buffer value determined based on business needs to obtain the minimum lower capacity limit and the maximum upper capacity limit. Combining the minimum lower capacity limit and the maximum upper capacity limit yields a capacity matching constraint interval. This interval reflects the minimum guaranteed demand and the maximum possible demand for materials within the demand time window, ensuring the matching process covers both normal and fluctuating demand scenarios. Subsequently, the material code of the first production demand material and the capacity matching constraint interval are used as search criteria to traverse the cross-border supplier qualification database. During this traversal, it is compared whether the supplier's qualification certification covers the material code, ensuring the candidate supplier possesses the legality and professionalism to supply goods. After passing the qualification certification comparison, the supplier's real-time capacity range and capacity elasticity limit are compared to determine whether it can meet the requirements of the capacity matching constraint interval. Only when a supplier's capacity elasticity limit covers or exceeds the upper limit of the capacity matching constraint interval is the supplier considered to have sufficient supply capacity to ensure a stable supply of materials under demand fluctuations. Through this process, K qualified suppliers that meet the criteria can be selected from the supplier qualification database. These qualified suppliers not only have the qualification certification for the first production demand materials, but their production capacity also fully covers the constraint range of the first time-series net demand in both time and quantity dimensions, which can ensure the resilience and controllability of the supply chain in subsequent links.

[0029] The task allocation module 13 is used to allocate production tasks based on the first time-series net demand and the K real-time capacity constraints of the K qualified suppliers, so as to obtain K time-series production task sequences.

[0030] Specifically, in the task allocation module 13, after identifying K qualified suppliers, the first time-series net demand is divided according to the demand time window, forming multiple discrete demand units. Each unit corresponds to a specific quantity requirement and delivery sequence. Subsequently, these demand units are dynamically mapped and compared with the K real-time capacity constraints of the K suppliers. Under the premise of ensuring that the total demand is met, tasks are allocated according to the preset optimization objectives to determine the optimal allocation ratio of each demand unit among the K suppliers. The allocation results are then mapped back to the time series to form K time-series production task sequences. Each time-series production task sequence not only clarifies the specific material production quantity that the supplier needs to complete at different time points, but also ensures the supply and demand balance of the overall supply chain and the timely progress of the production plan, thereby providing a data foundation for subsequent production redundancy prediction and transportation compensation.

[0031] Furthermore, the task allocation module 13 also includes:

[0032] Based on the time-series demand window, the first time-series net demand is divided into multiple discrete demand units, and a demand time-slice mapping table is constructed. K real-time capacity constraints from the K qualified suppliers are retrieved. Based on the time-series demand window, the K real-time capacity constraints are divided into K groups of time-series capacity elasticity intervals. The K groups of time-series capacity elasticity intervals are dynamically bound to the time-slice mapping table according to the time-slice dimension, constructing a three-dimensional constraint matrix. On the three-dimensional constraint matrix, with the goal of minimizing total delay and capacity overflow, time-series collaborative optimization of production tasks is performed to generate a supply decision matrix. Based on the K qualified suppliers, the supply decision matrix is ​​sliced, and the production tasks are reorganized in ascending order to output the K time-series production task sequences.

[0033] In one feasible implementation, the first time-series net demand is divided into several discrete demand units based on the time-series demand window. Each discrete demand unit contains a specific material quantity and a required delivery time. By mapping each discrete demand unit to a time slice, a demand time slice mapping table is constructed, providing a framework for subsequent capacity constraint binding. Subsequently, based on the unique identifiers of K qualified suppliers, K real-time capacity constraint information is extracted from the cross-border supplier qualification database. These real-time capacity constraints include not only the supplier's producible quantity within a specific time period but also its capacity adjustment capability under load fluctuations. Then, based on the time-series demand window, each supplier's real-time capacity constraint is divided into multiple segments corresponding to the demand time slice, forming K sets of time-series capacity elasticity intervals. In this way, each demand time slice can be matched with the corresponding capacity interval, ensuring the consistency of supply and demand data in the time dimension. After obtaining the K sets of time-series capacity elasticity intervals, these time-series capacity elasticity intervals are bound to the demand time slice mapping table one by one according to the time slice dimension, thereby constructing a three-dimensional constraint matrix to represent the quantity relationship and time constraints in the supply and demand matching process. Then, within the three-dimensional constraint matrix, time-series collaborative optimization of production tasks is performed with the goal of minimizing total delay and capacity overflow. Specifically, a multi-objective function is first constructed with the core objective of minimizing total delay penalty cost and capacity overflow loss. The total delay penalty is calculated by multiplying the deviation of the actual allocation time from the demand time window by a dynamic delay penalty factor, and the capacity overflow loss is quantified by multiplying the supplier's idle capacity by the opportunity cost per unit capacity. Subsequently, hard boundary conditions are loaded from the three-dimensional constraint matrix, including that the allocation amount of each supplier in each time slot must be within its capacity elasticity range, the sum of the allocation amounts of all suppliers in each time slot must be strictly equal to the demand in that time slot, and the fluctuation of the allocation amount in adjacent time slots must not exceed a preset stability threshold. Then, based on the simulated annealing optimization engine, a neighborhood search is performed in the solution space. New solutions are generated through three types of operations: supplier replacement, time slot shifting, and allocation amount fine-tuning. The new solutions are accepted based on the changes in the multi-objective weighted score and the probability of annealing temperature. The initial temperature of the simulated annealing optimization engine can be set based on the historical cost average, and the cooling strategy can adopt exponential annealing. After a predetermined number of iterations, the solution converges to the Pareto optimal set. The solution with the lowest total cost is selected and encapsulated as the supply decision matrix. This supply decision matrix stores the optimal allocation for each supplier in each time slice in the form of a three-dimensional tensor. Finally, based on the supply decision matrix of K qualified suppliers, each slice is processed, and the production tasks of each supplier are reorganized in ascending time order to obtain K time-series production task sequences. These time-series production task sequences not only ensure the matching of demand in terms of time and quantity but also balance the capacity utilization among suppliers, improving the overall execution efficiency and stability of the cross-border supply chain.

[0034] The production compensation module 14 is used to predict the production redundancy buffer of the K time-series production tasks based on the material production attributes of the K qualified suppliers, and to compensate the production process based on the prediction results to obtain K compensated time-series production redundancies.

[0035] Specifically, in the production compensation module 14, after obtaining the time-series production task sequences of each qualified supplier, the system calls upon the material production attributes corresponding to each supplier, such as the yield fluctuation coefficient and common production delay probabilities in the supplier's historical production process. Subsequently, a probabilistic simulation method (such as Monte Carlo simulation) is introduced to predict production redundancy buffers based on the material production attributes corresponding to each supplier, and to deduce the differences between actual and planned output under multiple possible production disturbance scenarios. Afterward, the production time windows of the time-series production task sequences are automatically compensated based on the simulation results to absorb possible fluctuation impacts, thereby generating K compensated time-series production redundancies for K suppliers. These compensated time-series production redundancies not only ensure that downstream demand can still be met under adverse conditions, but also provide a guarantee for the stable connection of cross-border transportation links.

[0036] Furthermore, the production compensation module 14 includes:

[0037] Retrieve K material production attributes from the K qualified suppliers, wherein the material production attributes include historical yield volatility coefficient and historical delay probability distribution; perform Monte Carlo simulation based on the K material production attributes to generate K production delay confidence intervals; use the K production delay confidence intervals to dynamically compensate the starting production time window of the K time-series production task sequences, and output the K compensated time-series production redundancy.

[0038] In a preferred embodiment, for K qualified suppliers, their corresponding material production attributes are retrieved one by one. These attributes include the supplier's historical yield volatility coefficient and historical delay probability distribution. The historical yield volatility coefficient reflects the stability of the supplier's yield across different batches; the historical delay probability distribution reflects the possible delay duration and its probability density within a given production cycle. Subsequently, based on probabilistic statistical methods, a hybrid event-triggered model suitable for each supplier is constructed, and extensive iterative simulations are performed using Monte Carlo simulation. In each simulation, random samples are taken from the supplier's historical yield volatility distribution and delay probability distribution to calculate the possible completion time of the task under the given disturbance conditions. Through repeated simulations, the system ultimately generates a delay duration dataset covering various uncertainties, and calculates the probability distribution interval of the delay duration based on this dataset, obtaining the production delay confidence interval for each supplier. These production delay confidence intervals clearly characterize the optimal, worst, and high-probability ranges of task completion time. Next, the production delay confidence intervals are matched with their respective time-sequential production task sequences. If a task has a high probability of delay risk within the original planned time window, the system automatically performs dynamic pre-compensation on the task's starting production time window based on the boundaries of the delay interval. In other words, additional redundant time is reserved before the plan is executed. In this way, the system can obtain K compensated time-sequential production redundancies. These redundancies not only ensure that production can be completed on time under fluctuating conditions, but also provide a guarantee for the connection of subsequent transportation links and the overall resilience of the cross-border supply chain.

[0039] Furthermore, the production compensation module 14 includes:

[0040] The system retrieves the first historical equipment failure rate from the first qualified supplier; constructs a hybrid event-triggered model based on the first historical good product volatility coefficient, the first historical delay probability distribution, and the first historical equipment failure rate; performs P Monte Carlo simulations on the hybrid event-triggered model to obtain a delay duration dataset; and cuts the delay distribution boundary according to the delay duration dataset to obtain the first production delay confidence interval.

[0041] In one feasible implementation, a first qualified supplier is randomly selected from K qualified suppliers. The first qualified supplier's historical equipment failure rate is retrieved from the cross-border supplier qualification database. This historical equipment failure rate characterizes the frequency and probability of equipment failure within a certain statistical period, and is a crucial factor affecting production progress and task completion time. Subsequently, the extracted historical equipment failure rate is combined with the first historical yield fluctuation coefficient and the first historical delay probability distribution to construct a hybrid event-triggered model. In this hybrid event-triggered model, yield fluctuation, delay probability, and equipment failure events are modeled as random variables, and their interactions are established by setting triggering conditions. For example, when the sampled yield rate is below a threshold, a production rework event is triggered; when the equipment failure rate variable falls into the failure range, an equipment downtime event is triggered; and when the delay distribution variable is in the tail probability range, a long-term delay event is triggered. Through these triggering conditions, the hybrid event-triggered model can realistically reproduce various disturbances that may occur in actual production. Subsequently, P Monte Carlo simulations are performed on the hybrid event-triggered model. In each simulation, values ​​for good product fluctuation, delay probability, and equipment failure are independently and randomly sampled, and the possible delay duration of the production task is calculated accordingly. After P iterations, a delay duration dataset is accumulated, which contains delay sample values ​​under various possible disturbance scenarios. Then, statistical analysis is performed on the obtained delay duration dataset. Based on a preset confidence level (e.g., 95% or 99%), the delay distribution is boundary-cut to obtain the upper and lower bounds of the delay duration, thus outputting the first production delay confidence interval. This first production delay confidence interval not only reflects the possible delay range of the first qualified supplier under historical attribute conditions, but also provides a quantitative basis for subsequent production redundancy compensation, ensuring that the cross-border supply chain has sufficient resilience and preventive capabilities in the production process.

[0042] The transportation compensation module 15 is used to predict cross-border transportation congestion for the K time-series production task sequences based on the material transportation attributes of the K qualified suppliers, and to perform transportation time-series compensation based on the prediction results to obtain K compensated time-series transportation tasks.

[0043] Specifically, in the transportation compensation module 15, the material transportation attributes of K qualified suppliers are extracted from the cross-border supplier qualification database. These material transportation attributes include the geographical location of the supplier's location, cross-border transportation routes, etc. Subsequently, based on these transportation attributes, the system constructs the topological structure of the cross-border transportation route, clarifying the origin, passing nodes, and destination of each task. After that, multi-source real-time logistics data streams are accessed, and these dynamic data are fused with the transportation route topology to predict the cross-border transportation congestion of K sequential production task sequences. According to the prediction results, the time windows of the K sequential production task sequences are dynamically corrected to obtain K compensated sequential transportation tasks. These compensated sequential transportation tasks not only ensure the timely delivery of cross-border materials under congested or adverse conditions but also provide proactive risk prevention and control capabilities for the transportation link of the overall supply chain's resilience strategy.

[0044] Furthermore, the transportation compensation module 15 includes:

[0045] Based on the K material transportation attributes of the K qualified suppliers, perform multimodal transportation path topology modeling to obtain K transportation path topologies; access K real-time multi-source logistics data streams, predict the transportation delay probability in the K transportation path topologies to obtain K delayed sequential path topologies; calculate the cumulative delay probability distribution based on the K delayed sequential path topologies to obtain K delay cumulative duration distributions; use the K delay cumulative duration distributions to dynamically compensate the start time windows of the K sequential production task sequences, and output the K compensated sequential transportation tasks.

[0046] In a preferred embodiment, the material transportation attributes of K qualified suppliers are first retrieved. These attributes include the supplier's geographical location, transportation mode, historical transportation timeliness, average customs clearance time, and transshipment node location. Based on this information, a multimodal transport modeling method is used to establish a transportation route topology for each supplier. In the transportation route topology, nodes represent transportation links, such as departure ports, land ports, transit hubs, and destination country customs clearance ports, while edges represent transportation routes and their transportation modes, with historical average duration and timeliness fluctuation distributions appended to the edges. Subsequently, the system accesses K real-time multi-source logistics data streams, which typically originate from real-time transportation monitoring systems, weather forecasting systems, traffic monitoring systems, etc., and these dynamic data are fused with the aforementioned transportation route topology. After obtaining these data streams, the corresponding data stream is overlaid on each route. Next, transportation, weather, and traffic data for each route, along with the relevant nodes, are input into a delay prediction model built on a long short-term neural network. This model predicts the delay probability of each transportation segment within a specific time window. The predicted delay probabilities are then synchronized to the corresponding transportation route topology, generating K time-series delay route topologies. Then, the joint probability distributions of delays for each segment are merged using probabilistic convolution to calculate the composite distribution of the total delay duration. Multiple Monte Carlo simulations are then performed on each route, and a cumulative delay duration dataset is generated based on the composite distribution. By calculating confidence intervals for the cumulative delay duration dataset, K cumulative delay duration distributions are obtained. Finally, the cumulative delay duration distributions are integrated with the corresponding time-series production task sequences. If the prediction indicates a high cumulative delay risk for a transportation route within the original departure time window, the departure time will be dynamically compensated based on the boundary values ​​of the distribution interval. This involves either appropriately advancing the departure time or configuring alternative transportation routes for the route. Through this compensation mechanism, an adjusted compensation-timed transportation task is ultimately output for each supplier, resulting in K compensation-timed transportation tasks. This effectively alleviates the risk of congestion in cross-border transportation and ensures that supply chain tasks can be delivered on schedule.

[0047] The compensation fusion module 16 is used to fuse the K compensation time-series production redundancies and K compensation time-series transportation tasks as the first cross-border supply chain resilience strategy.

[0048] Specifically, in the compensation fusion module 16, after obtaining K compensation-sequence production redundancies and K compensation-sequence transportation tasks, conflict detection is performed on the two types of compensation results. The system analyzes whether there is a contradiction between the task time window adjusted by the production redundancy and the departure time window moved forward or backward by the transportation compensation. For example, if the production redundancy of a certain material requires postponing the production end time, but the transportation compensation requires moving the departure time forward, the system will identify it as a conflict node. After detecting the conflict node, adjustments are made according to a priority strategy to resolve the conflict, generating a first cross-border supply chain resilience strategy covering both production and transportation. This improves the system's responsiveness and stability in the face of uncertainties such as production fluctuations and transportation congestion, thereby ensuring the safe and resilient operation of the entire cross-border supply chain.

[0049] Furthermore, the compensation fusion module 16 includes:

[0050] Spatiotemporal conflict detection is performed on the K compensated time-series production redundancies and K compensated time-series transportation tasks to obtain K sets of conflict nodes; based on the K sets of conflict nodes, the production windows of the K compensated time-series production redundancies are shifted forward, and the K compensated time-series production redundancies and K compensated time-series transportation tasks are merged to output the first cross-border supply chain resilience strategy.

[0051] In a preferred embodiment, during the fusion process, the production compensation time window for each material under the same supplier to compensate for production redundancy is first mapped to the transportation compensation time window for transportation tasks to compensate for production redundancy. The logical sequence of production completion and transportation start times is compared to identify any inconsistencies where production is not yet complete but transportation must begin. This results in the output of K conflict nodes, which accurately mark the locations of material tasks with timing conflicts. Subsequently, a preset priority strategy is read, and the time windows are adjusted accordingly. Specifically, the priority strategy prioritizes ensuring overall delivery timeliness by advancing the compensation time windows for the K production redundancy compensation tasks, shortening the production buffer time, ensuring that production can complete delivery preparation before transportation starts, and eliminating time conflicts between production and transportation. Subsequently, the revised K compensation time-series production redundancies and K compensation time-series transportation tasks are re-bound and time-aligned to form a conflict-free first cross-border supply chain resilience strategy covering both production and transportation. This first cross-border supply chain resilience strategy includes optimized production start and end times, transportation start times, and redundancy compensation windows, which can maintain the overall coordination and continuity of the cross-border supply chain under multi-source risk disturbances, providing enterprises with an actionable proactive risk avoidance solution.

[0052] Risk avoidance module 17 is used to construct a set of cross-border supply chain resilience strategies covering all material needs of production based on the multiple time-series net demand quantities of the various production demand materials, and to implement proactive risk avoidance in all aspects of the cross-border supply chain.

[0053] Specifically, in risk mitigation module 17, the time-series net demand for other production materials is analyzed using the same method described above. Cross-border supply chain resilience strategies for each production material are then derived and aggregated to form a comprehensive set of strategies covering all material levels. This set encompasses the entire process from raw material procurement, cross-border transportation, intermediate component processing, to final product delivery. Through this set of strategies, the system can identify the time and resource dependencies between different materials and, upon the occurrence of risk predictions, automatically select the optimal redundancy scheme and compensation path for proactive risk mitigation. This ensures the continuity and stability of the overall production chain, preventing actual interruptions and achieving dynamic, stable operation and resilience management across the entire cross-border supply chain.

[0054] In summary, the risk prediction and management system for cross-border supply chains provided in this application has the following technical effects:

[0055] The plan decomposition module 11 is used to download the master production plan from the order management terminal and decompose the master production plan based on the bill of materials hierarchy to obtain multiple time-series net requirements for various production demand materials; the supplier retrieval module 12 is used to use the first production demand material and the first time-series net requirement as joint retrieval conditions to traverse the cross-border supplier qualification database to obtain K qualified suppliers; the task allocation module 13 is used to allocate production tasks based on the first time-series net requirement and the K real-time capacity constraints of the K qualified suppliers to obtain K time-series production task sequences; the production compensation module 14 is used to perform production redundancy of the K time-series production task sequences based on the material production attributes of the K qualified suppliers. The system employs a buffer prediction mechanism and compensates for production process delays based on the prediction results, resulting in K compensated time-series production redundancies. A transportation compensation module 15 predicts cross-border transportation congestion for the K time-series production tasks based on the material transportation attributes of the K qualified suppliers, and compensates for transportation delays based on the prediction results, resulting in K compensated time-series transportation tasks. A compensation fusion module 16 fuses the K compensated time-series production redundancies and K compensated time-series transportation tasks as a first cross-border supply chain resilience strategy. A risk avoidance module 17 constructs a set of cross-border supply chain resilience strategies covering all material needs of production, based on multiple time-series net demand quantities of various production demand materials, and executes proactive risk avoidance across the entire cross-border supply chain. Through these steps, the system addresses the technical problem of insufficient supply chain resilience caused by multi-source uncertainties in the production and transportation stages of the cross-border supply chain, making it difficult to achieve proactive risk avoidance across all stages. It achieves the technical effect of improving the proactive risk avoidance capability and resource optimization efficiency of the cross-border supply chain, and reducing delivery delays and operating costs by dynamically predicting and compensating for production redundancy and transportation congestion.

[0056] Example 2, based on the same inventive concept as the risk prediction and management system for cross-border supply chains in the foregoing examples, such as... Figure 2 As shown in the embodiments of this application, a risk prediction and management method for cross-border supply chains is provided, the method comprising:

[0057] Download the master production plan from the order management terminal and decompose the master production plan based on the bill of materials hierarchy to obtain multiple time-series net requirements for various production demand materials; use the first production demand material and the first time-series net requirement as joint search conditions to traverse the cross-border supplier qualification database and obtain K qualified suppliers; based on the first time-series net requirement and the K real-time capacity constraints of the K qualified suppliers, allocate production tasks to obtain K time-series production task sequences; predict the production redundancy buffer for the K time-series production task sequences based on the material production attributes of the K qualified suppliers, and... Based on the prediction results, production process compensation is performed to obtain K compensated time-series production redundancies; based on the material transportation attributes of the K qualified suppliers, cross-border transportation congestion prediction is performed for the K time-series production task sequences, and transportation time-series compensation is performed based on the prediction results to obtain K compensated time-series transportation tasks; the K compensated time-series production redundancies and K compensated time-series transportation tasks are integrated as the first cross-border supply chain resilience strategy; by analogy, based on the multiple time-series net demand quantities of the various production demand materials, a cross-border supply chain resilience strategy set covering the demand for all production materials is constructed to implement proactive risk avoidance in all aspects of the cross-border supply chain.

[0058] Furthermore, the methods include:

[0059] The master production schedule is decomposed to obtain the time-series production quantity and time-series demand window of the target production product; the bill of materials is loaded according to the product code of the target production product, and a material hierarchy dependency relationship is constructed based on the multi-level nested structure of the bill of materials; based on the material hierarchy dependency relationship, the material demand back offset calculation of the time-series production quantity and time-series demand window is performed to obtain multiple time-series net demand quantities of the various production demand materials.

[0060] Furthermore, the methods include:

[0061] Extract the first production quantity and the first demand time window from the time-series production quantity and time-series demand time window mapping; calculate the material demand reverse offset of the first production quantity based on the material hierarchy dependency relationship to obtain the first material gross demand; aggregate the first material gross demand to obtain multiple first net demands of the various production demand materials; compensate the first demand time window in advance according to the standard production time to output the first production time window; bind the first production time window to the multiple first net demands to obtain the first time-series net demand data group; after obtaining multiple time-series net demand data groups corresponding to the time-series demand time windows by analogy, reorganize the multiple time-series net demand data groups based on the various production demand materials to obtain multiple time-series net demands of the various production demand materials.

[0062] Furthermore, the methods include:

[0063] Extract the capacity matching constraint interval from the first time-series net demand; use the material code of the first production demand material and the first net demand interval as joint search conditions to traverse the cross-border supplier qualification database to obtain the K qualified suppliers; wherein, the qualification certification of the qualified suppliers includes the first production demand material, and the upper limit of capacity elasticity covers the capacity matching constraint interval.

[0064] Furthermore, the methods include:

[0065] Based on the time-series demand window, the first time-series net demand is divided into multiple discrete demand units, and a demand time-slice mapping table is constructed. K real-time capacity constraints from the K qualified suppliers are retrieved. Based on the time-series demand window, the K real-time capacity constraints are divided into K groups of time-series capacity elasticity intervals. The K groups of time-series capacity elasticity intervals are dynamically bound to the time-slice mapping table according to the time-slice dimension, constructing a three-dimensional constraint matrix. On the three-dimensional constraint matrix, with the goal of minimizing total delay and capacity overflow, time-series collaborative optimization of production tasks is performed to generate a supply decision matrix. Based on the K qualified suppliers, the supply decision matrix is ​​sliced, and the production tasks are reorganized in ascending order to output the K time-series production task sequences.

[0066] Furthermore, the methods include:

[0067] Based on the K material transportation attributes of the K qualified suppliers, multimodal transport route topology modeling is performed to obtain K transport route topologies; K real-time multi-source logistics data streams are accessed, and transport delay probability prediction is performed on the K transport route topologies to obtain K delay time-series route topologies; based on the cumulative delay probability distribution of the K delay time-series route topologies, K cumulative delay duration distributions are calculated; the K cumulative delay duration distributions are used to dynamically compensate the departure time windows of the K time-series production task sequences, and the K compensated time-series transport tasks are output.

[0068] Furthermore, the methods include:

[0069] Retrieve K material production attributes from the K qualified suppliers, wherein the material production attributes include historical yield volatility coefficient and historical delay probability distribution; perform Monte Carlo simulation based on the K material production attributes to generate K production delay confidence intervals; use the K production delay confidence intervals to dynamically compensate the starting production time window of the K time-series production task sequences, and output the K compensated time-series production redundancy.

[0070] Furthermore, the methods include:

[0071] The system retrieves the first historical equipment failure rate from the first qualified supplier; constructs a hybrid event-triggered model based on the first historical good product volatility coefficient, the first historical delay probability distribution, and the first historical equipment failure rate; performs P Monte Carlo simulations on the hybrid event-triggered model to obtain a delay duration dataset; and cuts the delay distribution boundary according to the delay duration dataset to obtain the first production delay confidence interval.

[0072] Furthermore, the methods include:

[0073] Spatiotemporal conflict detection is performed on the K compensated time-series production redundancies and K compensated time-series transportation tasks to obtain K sets of conflict nodes; based on the K sets of conflict nodes, the production windows of the K compensated time-series production redundancies are shifted forward, and the K compensated time-series production redundancies and K compensated time-series transportation tasks are merged to output the first cross-border supply chain resilience strategy.

[0074] In summary, any step of the method described above can be stored as a computer instruction or program in an unrestricted computer memory, and can be called and identified by an unrestricted computer processor to implement any method in the embodiments of this application, without any additional restrictions.

[0075] Furthermore, the "first" or "second" mentioned above may not only represent a sequential relationship, but may also represent a specific concept, and / or refer to the individual or collective selection of multiple elements. Clearly, those skilled in the art can make various modifications and variations to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A risk prediction and management system for cross-border supply chains, characterized in that: The system includes: The planning decomposition module is used to download the master production plan from the order management terminal and decompose the master production plan based on the bill of materials hierarchy to obtain multiple time-sequential net requirements for various production demand materials; The supplier search module is used to combine the first production demand material and the first time-series net demand as joint search conditions, traverse the cross-border supplier qualification database, and obtain K qualified suppliers. The task allocation module is used to allocate production tasks based on the first time-series net demand and the K real-time capacity constraints of the K qualified suppliers, so as to obtain K time-series production task sequences. The production compensation module is used to predict the production redundancy buffer of the K time-series production tasks based on the material production attributes of the K qualified suppliers, and to compensate the production process based on the prediction results to obtain K compensated time-series production redundancies. The transportation compensation module is used to predict cross-border transportation congestion for the K time-series production tasks based on the material transportation attributes of the K qualified suppliers, and to perform transportation time-series compensation based on the prediction results to obtain K compensated time-series transportation tasks. The compensation fusion module is used to fuse the K compensation time-series production redundancies and K compensation time-series transportation tasks as the first cross-border supply chain resilience strategy. The risk avoidance module is used to construct a set of cross-border supply chain resilience strategies covering all material needs of production based on the multiple time-series net demand quantities of the various production demand materials, and to implement proactive risk avoidance in all aspects of the cross-border supply chain.

2. The risk prediction and management system for cross-border supply chains as described in claim 1, characterized in that, The plan breakdown module includes: Decompose the master production schedule to obtain the time-series production quantity and time-series demand window for the target product. Load the bill of materials based on the product code of the target product, and construct the material hierarchy dependency relationship based on the multi-level nested structure of the bill of materials; Based on the material hierarchy dependency relationship, the material demand back offset calculation is performed on the time-series production quantity and the time-series demand time window to obtain multiple time-series net demand quantities of the various production demand materials.

3. The risk prediction and management system for cross-border supply chains as described in claim 2, characterized in that, The plan breakdown module includes: Extract the first production quantity and the first demand time window from the time-series production quantity and time-series demand time window mapping; Based on the material hierarchy dependency relationship, the material requirement reverse offset calculation for the first production quantity is performed to obtain the first material gross requirement; Aggregate the gross demand of the first material to obtain multiple first net demands of the various production demand materials; Based on the standard production time, the first demand time window is pre-compensated, and the first production time window is output. Bind the first production time window to the plurality of first net demand quantities to obtain the first time-series net demand data group; After obtaining multiple time-series net demand data sets corresponding to the time-series demand window by analogy, the multiple time-series net demand data sets are reorganized based on the multiple production demand materials to obtain multiple time-series net demand quantities of the multiple production demand materials.

4. The risk prediction and management system for cross-border supply chains as described in claim 1, characterized in that, The supplier search module includes: Extract the capacity matching constraint interval from the first time-series net demand; Using the material code of the first production demand material and the first net demand range as joint search conditions, the cross-border supplier qualification database is traversed to obtain the K qualified suppliers. The qualification certification of the qualified supplier includes the first production demand material, and the upper limit of the capacity flexibility covers the capacity matching constraint range.

5. The risk prediction and management system for cross-border supply chains as described in claim 2, characterized in that, The task allocation module includes: Based on the time window of the time-series demand, the first time-series net demand is divided into multiple discrete demand units, and a demand time slice mapping table is constructed. Retrieve the K real-time capacity constraints of the K qualified suppliers; Based on the time window of the time sequence demand, the K real-time capacity constraints are divided into K groups of time sequence capacity elasticity intervals; The K groups of time-series capacity elasticity intervals are dynamically bound to the time-slice mapping table according to the time-slice dimension to construct a three-dimensional constraint matrix; Based on the three-dimensional constraint matrix, with the goal of minimizing total delay and capacity overflow, a production task timing collaborative optimization is performed to generate a supply decision matrix; Based on the K qualified suppliers, the supply decision matrix is ​​sliced, and the production tasks are reorganized in ascending order of time to output the K time-series production tasks.

6. The risk prediction and management system for cross-border supply chains as described in claim 1, characterized in that, The transportation compensation module includes: Based on the K material transportation attributes of the K qualified suppliers, multimodal transport route topology modeling is performed to obtain K transport route topologies; By accessing K real-time multi-source logistics data streams, and performing transportation delay probability prediction on the K transportation path topologies, K delay time-series path topologies are obtained; Based on the cumulative delay probability distribution of the K delayed time sequence paths, the cumulative delay duration distribution of the K delays is obtained; The K cumulative delay durations are used to dynamically compensate the start time windows of the K time-series production tasks, and the K compensated time-series transportation tasks are output.

7. The risk prediction and management system for cross-border supply chains as described in claim 1, characterized in that, The production compensation module includes: Retrieve K material production attributes from the K qualified suppliers, wherein the material production attributes include historical yield volatility coefficient and historical delay probability distribution; Based on the K material production attributes, a Monte Carlo simulation is performed to generate K production delay confidence intervals; The K production delay confidence intervals are used to dynamically compensate the starting production time windows of the K time-series production task sequences, and the K compensated time-series production redundancies are output.

8. The risk prediction and management system for cross-border supply chains as described in claim 7, characterized in that, The production compensation module includes: The first historical equipment failure rate of the first qualified supplier was retrieved; Based on the first historical good product volatility coefficient, the first historical delay probability distribution, and the first historical equipment failure rate, a hybrid event triggering model is constructed. P Monte Carlo simulations were performed on the hybrid event-triggered model to obtain a dataset of delay durations. Based on the delay duration dataset, the delay distribution boundary is cut to obtain the first production delay confidence interval.

9. The risk prediction and management system for cross-border supply chains as described in claim 1, characterized in that, The compensation and fusion module includes: Spatiotemporal conflict detection is performed on the K compensated time-series production redundancy and K compensated time-series transportation tasks to obtain K sets of conflict nodes; Based on the K groups of conflict nodes, after shifting the production windows of the K compensated time-series production redundancies forward, the K compensated time-series production redundancies and K compensated time-series transportation tasks are integrated to output the first cross-border supply chain resilience strategy.

10. A risk prediction and management method for cross-border supply chains, characterized in that: The method is executed through a risk prediction and management system for cross-border supply chains as described in any one of claims 1 to 9, the method comprising: Download the master production plan from the order management terminal, and decompose the master production plan based on the bill of materials hierarchy to obtain multiple time-sequential net requirements for various production demand materials; Using the first production demand material and the first time-series net demand as joint search conditions, we traversed the cross-border supplier qualification database to obtain K qualified suppliers. Based on the first time-series net demand and the K real-time capacity constraints of the K qualified suppliers, production tasks are allocated to obtain K time-series production task sequences. Based on the material production attributes of the K qualified suppliers, the production redundancy buffer prediction of the K time-series production task sequences is performed, and the production process is compensated based on the prediction results to obtain K compensated time-series production redundancies. Based on the material transportation attributes of the K qualified suppliers, cross-border transportation congestion is predicted for the K time-series production tasks, and transportation time-series compensation is performed based on the prediction results to obtain K compensated time-series transportation tasks. The K compensated time-series production redundancies and K compensated time-series transportation tasks are integrated as the first cross-border supply chain resilience strategy; By analogy, based on the multiple time-series net demand quantities of various production demand materials, a set of cross-border supply chain resilience strategies covering the demand for all production materials is constructed to proactively mitigate risks throughout the entire cross-border supply chain.

Images