Warehousing network model and algorithm based on two-stage model collaborative optimization

Abstract

The invention discloses a warehousing network model and algorithm based on two-stage model collaborative optimization, and belongs to the technical field of warehousing network models. The invention discloses a warehousing network model and algorithm based on two-stage model collaborative optimization. The warehousing network model comprises an analysis and prediction module, an inventory collaborative management module and an optimization model. According to the invention, the problem of lack of overall planning of the full-link cost of the storage network in the prior art is solved. Through collaborative optimization of the first stage model and the second stage model, balance of long-term cost and short-term cost is realized, full-link cost optimization is achieved, the storage space utilization rate is improved, maximum utilization of manpower, equipment and space resources is realized, and through integration of regional economic data, population density, a traffic network, competitor layout and other multi-dimensional data, the real-time performance of the system is improved. Costs, efficiency and risks of different site selection schemes are calculated, a clear quantitative basis is provided for long-term decisions such as storage node layout and facility scale determination, and large deviation of demand probability fitting is avoided.

Description

Technical Field

[0001] This invention relates to the field of warehouse network modeling technology, specifically to a warehouse network model and algorithm based on two-stage model collaborative optimization. Background Technology

[0002] In the field of modern warehouse management, the rapid development and widespread application of data-driven technologies demonstrate their growing trend and enormous potential. With the continuous advancement of information technology and the improvement of data acquisition capabilities, warehouse management systems can collect, organize, and analyze large amounts of real-time and historical data, not only sales, inventory, and cost data, but also complex multi-source data such as supply chain information. This data is used to build accurate predictive models and optimization strategies, effectively improving warehouse efficiency, reducing costs, optimizing inventory management, and significantly enhancing supply chain responsiveness.

[0003] Chinese patent CN118735413B discloses a method and system for determining a warehouse management model. The method includes the following steps: acquiring inventory data; preprocessing the inventory data to generate warehouse management data; performing statistical analysis on the warehouse management data to generate warehouse statistical data; collecting warehouse statistical data in real time to obtain real-time warehouse management data; acquiring sales cost data and historical sales data; extracting numerical features from the sales cost data to generate warehouse numerical feature data; and calculating the product inventory turnover rate from the warehouse numerical feature data to generate overall turnover rate data. The timely feedback mechanism of this patent helps improve the flexibility and responsiveness of warehouse operations, enabling managers to respond more quickly to market changes and customer needs.

[0004] In practical use, the aforementioned patents improve the flexibility and responsiveness of warehouse operations through a timely feedback mechanism. While this mechanism can reduce short-term waste through real-time adjustments, it lacks comprehensive planning for the entire cost chain of the warehousing network; therefore, it does not meet current needs. To address this, we propose a warehousing network model and algorithm based on a two-stage model for collaborative optimization. Summary of the Invention

[0005] The purpose of this invention is to provide a warehouse network model and algorithm based on a two-stage model collaborative optimization. By collaboratively optimizing the Phase 1 and Phase 2 models, a balance between long-term and short-term costs is achieved, resulting in optimal end-to-end costs, improved warehouse space utilization, and maximized use of human, equipment, and space resources. The collaborative optimization of the Phase 1 and Phase 2 models significantly enhances network robustness, avoids supply chain disruptions caused by single node failures, and ensures network continuity. By integrating multi-dimensional data such as regional economic data, population density, transportation networks, and competitor layouts, the cost, efficiency, and risk of different site selection schemes are calculated, providing clear quantitative basis for long-term decisions such as warehouse node layout and facility scale determination. This avoids significant deviations in demand probability fitting and solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a warehouse network model based on two-stage model collaborative optimization, comprising:

[0007] The analysis and forecasting module is used to determine the optimal geographical locations of factory, transit station and warehouse nodes through spatial analysis, perform dynamic demand forecasting, and dynamically adjust the storage capacity of each node based on the forecast results.

[0008] The inventory collaborative management module is used to determine warehouse location decisions and pre-stocking decisions, and to calculate the objective function, material supply warehouse location constraints, warehouse capacity constraints for Phase 1, and network node flow balance constraints, network arc capacity constraints, and decision variable constraints for Phase 2.

[0009] The optimization module is used to plan the scheduling and allocation of emergency supplies in the warehousing network based on the location decision and pre-reservation decision results in Phase 1. It introduces intermediate variables to transform the Phase 1 model into a scenario probability constraint model and performs dual transformation.

[0010] Preferably, the analysis and prediction module includes:

[0011] The historical data analysis unit is used to collect past order data and market sales data, and to use time series analysis to analyze the changing trends of demand over time, seasonal fluctuation patterns, and their correlation with other variables.

[0012] The forecasting unit is used to dynamically forecast demand by combining demand trends over time, seasonal fluctuations, correlations with other variables, industry dynamics, and changes in consumer preferences. Based on the results of the dynamic demand forecast, the storage capacity of each node in the warehouse is adjusted.

[0013] Preferably, the inventory collaborative management module includes:

[0014] The site selection decision unit is used to determine the warehouse area and floor height parameters based on the storage capacity of each node in the warehouse, and to obtain key constraints and target priorities from five dimensions to determine the warehouse site selection decision.

[0015] The pre-reservation decision unit is used to clarify the decision boundary and input conditions. Combining the two-stage collaborative logic, it performs calculations according to the steps of planning rough calculation, operation fine calculation and collaborative optimization. It is also classified and stratified according to cargo type, node function and scenario requirements to obtain the pre-reservation decision.

[0016] The calculation unit is used to calculate the objective function, material supply warehouse location constraints, and warehouse capacity constraints for Phase 1, as well as the network node flow balance constraints, network arc capacity constraints, and decision variable constraints for Phase 2.

[0017] Preferably, the pre-reservation decision unit specifically includes:

[0018] Obtain key information across five dimensions: demand, supply, network planning, cost, and external risks, and clarify decision-making boundaries and input conditions;

[0019] Based on the decision boundary and input conditions, and combined with the two-stage collaborative logic, the calculation is performed according to the steps of rough planning calculation, operational fine calculation and collaborative optimization to obtain the specific value of the pre-reserve.

[0020] Based on the calculation results, the cargo type, node function, and scenario requirements are combined to classify and stratify the cargo to obtain the pre-storage decision.

[0021] Preferably, the computing unit specifically includes:

[0022] The Phase 1 model is used to calculate the objective function, material supply warehouse location constraints, and warehouse capacity constraints for Phase 1.

[0023] The Phase 2 model is used to calculate the network node flow balance constraints in Phase 2, as well as the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

[0024] Preferably, the stage one model includes:

[0025] The objective function for Phase 1 is calculated by minimizing the supply warehouse location cost, the emergency material pre-stocking procurement cost, and the expected cost under the worst-case demand distribution in Phase 2 after the location and pre-stocking quantities are determined. The calculation formula is as follows:

[0026]

[0027] In the formula, For site selection decisions, This is the transportation cost coefficient. For pre-reserve quantity decision, To determine the expected costs resulting from the pre-reservation and site selection decisions under the worst-case demand distribution in Stage 2, This is the inventory holding cost coefficient;

[0028] Calculate the site selection constraints for material supply warehouses, namely, the number of supply points selected at each node must not exceed 1;

[0029] Calculate warehouse capacity constraints, namely, the amount of pre-stored materials must not exceed the corresponding capacity of the supply warehouse and decision variable constraints.

[0030] Preferably, the stage two model includes:

[0031] The network node flow balance constraint in Phase 2 is calculated by subtracting the outflow and demand of each node from the sum of the inflow and pre-stored resources. If the difference is positive, the node will have surplus resources; otherwise, a negative difference indicates a shortage of resources.

[0032] Calculate the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

[0033] Preferably, the optimization module includes:

[0034] The allocation unit is used to plan and allocate emergency supplies in the warehousing network based on the site selection and pre-stocking decisions made in Phase 1.

[0035] The optimization algorithm unit is used to introduce intermediate variables to transform the stage one model into a scenario probability constraint model, transform the scenario probability constraint model into a dual problem, and after the dual transformation, transform the stage two model into a linear optimization model and calculate the dual problem with respect to uncertain scenario probabilities.

[0036] Preferably, the step of introducing intermediate variables to transform the stage-one model into a scenario probability constraint model specifically includes:

[0037] By substituting intermediate variables into the objective function of the stage one model, replacing the original fixed parameters, and incorporating the probability of scenario occurrence, the objective function is transformed from minimizing deterministic costs to minimizing expected costs that take into account scenario probabilities.

[0038] The deterministic constraints of the Phase 1 model are transformed into constraints that change dynamically with the scenario and meet probability requirements through intermediate variables.

[0039] The transformed scenario probability constraint model is solved using an algorithm, and the rationality of the intermediate variables and the robustness of the model are verified.

[0040] A two-stage model-based collaborative optimization algorithm is applied to a warehouse network model based on a two-stage model collaborative optimization, including the following steps:

[0041] S1: Determine the optimal geographical locations of factories, transit stations, and warehouse nodes through spatial analysis, perform dynamic demand forecasting, and dynamically adjust the storage capacity of each node based on the forecast results;

[0042] S2: Obtain key constraints and target priorities, determine warehouse location decisions, and classify them in layers based on cargo type, node function and scenario requirements to obtain pre-storage decisions;

[0043] S3: Calculate the objective function, material supply warehouse location constraints, warehouse capacity constraints, and decision variable constraints for Phase 1;

[0044] S4: Calculate the network node flow balance constraints in Phase 2, the network arc capacity constraints to be satisfied during transportation after a sudden event, and the constraints of the calculation decision variables;

[0045] S5: Based on the site selection and pre-stocking decisions in Phase 1, plan the scheduling and allocation of emergency supplies in the warehousing network, and calculate the dual problem regarding the probability of uncertain scenarios.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] This invention achieves a balance between long-term and short-term costs through the collaborative optimization of the Phase 1 and Phase 2 models, resulting in optimal end-to-end costs. Phase 1 analyzes long-term demand trends through demand forecasting to determine site selection and pre-stocking decisions. Phase 2, based on the site selection and pre-stocking decisions from Phase 1, plans the allocation and scheduling of emergency supplies within the warehousing network, avoiding over-construction of facilities, improving warehousing space utilization, and maximizing the use of human, equipment, and space resources. The collaborative optimization of the Phase 1 and Phase 2 models significantly enhances network robustness, preventing supply chain disruptions caused by single-node failures and ensuring network continuity. Phase 2 possesses real-time perception and rapid response capabilities, avoiding order delays or inventory backlogs caused by the lag in traditional model decisions, ensuring unimpeded fulfillment efficiency. By integrating multi-dimensional data such as regional economic data, population density, transportation networks, and competitor layouts, the cost, efficiency, and risk of different site selection schemes are calculated, providing clear quantitative basis for long-term decisions such as warehousing node layout and facility scale determination, avoiding significant deviations in demand probability fitting. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the warehouse network model module based on two-stage model collaborative optimization of the present invention;

[0049] Figure 2 This is a schematic diagram of the warehouse network model and algorithm based on two-stage model collaborative optimization of the present invention. Detailed Implementation

[0050] 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.

[0051] To address the issue of existing technologies using timely feedback mechanisms to improve warehouse operational flexibility and responsiveness, while these mechanisms can reduce short-term waste through real-time adjustments, they lack a comprehensive plan for the entire cost chain of the warehousing network. Please refer to [link to relevant documentation]. Figures 1-2 This embodiment provides the following technical solution:

[0052] A warehouse network model based on two-stage collaborative optimization includes:

[0053] The analysis and forecasting module is used to determine the optimal geographical location of factories, transit stations and warehouse nodes through spatial analysis. It comprehensively considers transportation networks, service radius and land costs to make dynamic demand forecasts and dynamically adjust the storage capacity of each node based on the forecast results to avoid resource idleness or overload.

[0054] The inventory collaborative management module is used to determine warehouse location decisions and pre-stocking decisions, and to calculate the objective function, material supply warehouse location constraints, warehouse capacity constraints for Phase 1, and network node flow balance constraints, network arc capacity constraints, and decision variable constraints for Phase 2.

[0055] The optimization module is used to plan the scheduling and allocation of emergency supplies in the warehousing network based on the location decision and pre-reservation decision results in Phase 1. It introduces intermediate variables to transform the Phase 1 model into a scenario probability constraint model and performs dual transformation.

[0056] The analysis and prediction module includes:

[0057] The historical data analysis unit is used to collect past order data and market sales data, and to use time series analysis to analyze the changing trends of demand over time, seasonal fluctuation patterns, and their correlation with other variables.

[0058] The forecasting unit is used to dynamically forecast demand by combining demand trends over time, seasonal fluctuations, correlations with other variables, industry dynamics, and changes in consumer preferences. Based on the results of the dynamic demand forecast, the storage capacity of each node in the warehouse is adjusted.

[0059] For example, by analyzing sales data of beverage products in summer over the years, and combining factors such as temperature and holidays, we can predict the demand scale of such products in different regions in the summer of the future. We can also pay attention to external factors such as industry dynamics, market research information, and changes in consumer preferences, and use expert judgment, scenario analysis and other means to predict the future trend of the market, thereby correcting the prediction results based on historical data.

[0060] By comprehensively analyzing various costs such as construction costs, operating costs, land costs, and labor costs, as well as the expected revenue growth and service level improvement after the warehouse is put into use, different site selection options are quantitatively evaluated. For example, by comparing the cost differences between building warehouses on the outskirts of the city and in the suburbs, and the impact on delivery timeliness and customer satisfaction, the optimal site is determined.

[0061] The inventory collaborative management module includes:

[0062] The location decision unit is used to determine the warehouse area and floor height parameters based on the storage capacity of each node in the warehouse, and to obtain key constraints and target priorities from five dimensions to determine the warehouse location decision. For example, it selects areas close to transportation hubs and with concentrated surrounding consumer markets as warehouse candidate locations to reduce transportation costs and respond quickly to customer needs. The core role of the location decision unit is to screen out the optimal warehouse location from the dimensions of geospatial and economic costs, build an efficient physical framework for the warehousing network, and balance transportation costs, service timeliness and operational feasibility.

[0063] The pre-reservation decision unit is used to clarify the decision boundary and input conditions. Combining the two-stage collaborative logic, it performs calculations according to the steps of planning rough calculation, operation fine calculation and collaborative optimization. It also performs hierarchical classification based on cargo type, node function and scenario requirements to obtain the pre-reservation decision.

[0064] The computational unit is used to calculate the objective function of Phase 1, the material supply warehouse location constraints, the warehouse capacity constraints, and the network node flow balance constraints, network arc capacity constraints, and decision variable constraints of Phase 2. It realizes the collaborative optimization of the two-stage model. The optimization objective of Phase 1 is composed of the objective of Phase 2, and the decision of Phase 2 is based on Phase 1. The dual problem about the probability of uncertain scenarios is solved by transforming the scenario probability constraint model into a dual problem.

[0065] The pre-reserve quantity decision-making unit specifically includes:

[0066] Obtain key information across five dimensions: demand, supply, network planning, cost, and external risks, and clarify decision-making boundaries and input conditions;

[0067] Based on the decision boundary and input conditions, and combined with the two-stage collaborative logic, the calculation is performed according to the steps of rough planning calculation, operational fine calculation and collaborative optimization to obtain the specific value of the pre-reserve.

[0068] Based on the calculation results, the cargo type, node function and scenario requirements are combined to classify and stratify the pre-storage volume decision.

[0069] The site selection decision unit specifically includes:

[0070] Define the function of the warehouse, and determine the warehouse area and floor height parameters;

[0071] Key constraints and priorities are derived from market demand, cost, transportation, policy, and operational suitability to avoid site selection that is detached from actual business scenarios.

[0072] By utilizing the key constraints and target priorities obtained, and combining the two-stage collaborative logic, the optimal location is selected from the candidate areas according to the steps of preliminary screening, quantitative evaluation, and collaborative optimization.

[0073] After determining the initial site selection, the number and scale of nodes in the warehouse network model and the transportation plan are combined for optimization to obtain the warehouse site selection decision, ensuring that the site selection is coordinated with the overall network.

[0074] The role of the inventory collaborative management module is to break down the information silos between modules. By constructing a unified objective function, setting constraints, and solving in stages, it achieves the overall optimization of the warehousing network, rather than the local optimization of a single module.

[0075] The computing unit specifically includes:

[0076] The Phase 1 model is used to calculate the objective function, material supply warehouse location constraints, and warehouse capacity constraints for Phase 1.

[0077] The Phase 2 model is used to calculate the network node flow balance constraints in Phase 2, as well as the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

[0078] Phase 1 model, including:

[0079] The objective function for Phase 1 is calculated by minimizing the supply warehouse location cost, the emergency material pre-stocking procurement cost, and the expected cost under the worst-case demand distribution in Phase 2 after the location and pre-stocking quantities are determined. The calculation formula is as follows:

[0080]

[0081] In the formula, For site selection decisions, This is the transportation cost coefficient. For pre-reserve quantity decision, To determine the expected costs resulting from the pre-reservation and site selection decisions under the worst-case demand distribution in Stage 2, This is the inventory holding cost coefficient;

[0082] Calculate the site selection constraints for material supply warehouses, namely, the number of supply points selected at each node must not exceed 1;

[0083] Calculate warehouse capacity constraints, namely, the amount of pre-stored materials must not exceed the corresponding capacity of the supply warehouse and decision variable constraints;

[0084] Phase 2 model, including:

[0085] The network node flow balance constraint in Phase 2 is calculated by subtracting the outflow and demand of each node from the sum of the inflow and pre-stored resources. If the difference is positive, the node will have surplus resources; otherwise, a negative value indicates a shortage of resources.

[0086] Calculate the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

[0087] The optimization module includes:

[0088] The allocation unit is used to plan and allocate emergency supplies in the warehousing network based on the site selection and pre-stocking decisions made in Phase 1.

[0089] The optimization algorithm unit is used to introduce intermediate variables to transform the stage one model into a scenario probability constraint model, transform the scenario probability constraint model into a dual problem, and after the dual transformation, transform the stage two model into a linear optimization model and calculate the dual problem with respect to uncertain scenario probabilities.

[0090] Introducing intermediate variables transforms the Phase 1 model into a scenario probability constraint model, specifically including:

[0091] By substituting intermediate variables into the objective function of the stage one model, replacing the original fixed parameters, and incorporating the probability of scenario occurrence, the objective function is transformed from minimizing deterministic costs to minimizing expected costs that take into account scenario probabilities.

[0092] The deterministic constraints of the Phase 1 model are transformed into constraints that change dynamically with the scenario and meet probability requirements through intermediate variables, ensuring that the model does not fail under high-probability scenarios while controlling low-probability risks.

[0093] The transformed scenario probability constraint model is solved by algorithm, and the rationality of intermediate variables and the robustness of the model are verified to avoid logical loopholes or parameter deviations during the transformation process.

[0094] Working principle: When using the warehouse network model and algorithm based on two-stage model collaborative optimization of this invention, according to... Figure 1 and Figure 2 This includes the following steps:

[0095] S1: Collect past order data and market sales data, use time series analysis to analyze the changing trend of demand over time, seasonal fluctuation patterns, and correlation with other variables, combine the changing trend of demand over time, seasonal fluctuation patterns, correlation with other variables, as well as industry dynamics and changes in consumer preferences, to make dynamic demand forecasts, and adjust the storage capacity of each node in the warehouse based on the results of dynamic demand forecasts.

[0096] S2: Determine the warehouse area and floor height parameters based on the storage capacity of each node in the warehouse, and obtain key constraints and target priorities from five dimensions to determine the warehouse location decision, clarify the decision boundary and input conditions, combine the two-stage collaborative logic, and perform calculations according to the planning rough calculation, operation fine calculation and collaborative optimization steps. Combine the cargo type, node function and scenario requirements to perform hierarchical classification and obtain the pre-storage decision.

[0097] S3: Calculate the objective function of Phase 1. The objective function of Phase 1 is a combination of minimizing the cost of supply warehouse location selection, the cost of emergency material pre-stocking procurement, and the expected cost brought about by the worst demand distribution in Phase 2 after the location and pre-stocking quantity are determined. Calculate the material supply warehouse location constraints, that is, the number of supply points selected at each node shall not be greater than 1. Calculate the warehouse capacity constraints, that is, the material pre-stocking quantity shall not exceed the corresponding capacity of the supply warehouse and the decision variable constraints.

[0098] S4: Calculate the network node flow balance constraint in Phase 2. The network node flow balance constraint in Phase 2 is calculated by subtracting the outflow and demand of each node from the sum of the inflow and pre-stored material of each node. If the difference is positive, the node will have surplus material; otherwise, a negative value indicates a material shortage. Calculate the network arc capacity constraint and decision variable constraint to be satisfied during transportation after a sudden event.

[0099] S5: Based on the site selection and pre-stocking decisions in Phase 1, schedule and allocate emergency supplies in the warehousing network. Introduce intermediate variables to transform the Phase 1 model into a scenario probability constraint model. Transform the scenario probability constraint model into a dual problem. The Phase 2 model is a nonlinear model. After dual transformation, transform the Phase 2 model into a linear optimization model and calculate the dual problem with respect to uncertain scenario probabilities.

[0100] In summary, this invention presents a warehouse network model and algorithm based on a two-stage collaborative optimization model. Stage 1 focuses on warehouse node location selection, facility scale determination, and resource allocation. The algorithm quantifies macroeconomic factors such as regional demand density, transportation costs, and land rent to avoid blind deployment or resource idleness. Stage 2 responds in real-time to order fluctuations, inventory levels, and transportation timeliness requirements, optimizing inventory allocation, replenishment routes, and order sorting processes. Stage 1 addresses the preventative warehouse location selection and material pre-stocking issues arising from uncertain demand distribution by analyzing long-term demand trends through demand forecasting, proactively deploying warehouse nodes in core areas to shorten the physical distance for order fulfillment. Stage 2, based on the location and pre-stocking decisions from Stage 1, plans the scheduling and allocation of emergency materials within the warehouse network. The Stage 1 and Stage 2 models achieve dynamic resource balance through a collaborative algorithm, determining reasonable warehouse facility capacity and avoiding over-construction. This is achieved through inventory sharing and cross-node adjustments. Optimization improves warehouse space utilization and maximizes the use of human, equipment, and space resources. The collaborative optimization of the Phase 1 and Phase 2 models significantly enhances network robustness through layered responses and dynamic adjustments. During the site selection phase, regional risks are comprehensively considered, and a redundant network is built through multi-node layout to avoid supply chain disruptions caused by single node failures, ensuring network continuity. Phase 2 has real-time perception and rapid response capabilities. When sudden demand fluctuations or supply chain disruptions occur, inventory allocation plans can be adjusted, transportation routes switched, or backup nodes activated within minutes. This avoids order delays or inventory backlogs caused by the lag in traditional model decisions, ensuring that fulfillment efficiency is not impacted. By integrating multi-dimensional data such as regional economic data, population density, transportation networks, and competitor layouts, a quantitative model is built to calculate the cost, efficiency, and risk of different site selection options. This provides a clear quantitative basis for long-term decisions such as warehouse node layout and facility scale determination, avoiding significant deviations in demand probability fitting.

[0101] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0102] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A warehouse network model based on two-stage collaborative optimization, characterized in that, include: The analysis and forecasting module is used to determine the optimal geographical locations of factories, transit stations, and warehouse nodes through spatial analysis, perform dynamic demand forecasting, and dynamically adjust the storage capacity of each node based on the forecast results. The inventory collaborative management module is used to determine warehouse location decisions and pre-stocking decisions, and to calculate the objective function, material supply warehouse location constraints, warehouse capacity constraints for Phase 1, and network node flow balance constraints, network arc capacity constraints, and decision variable constraints for Phase 2. The optimization module is used to plan the scheduling and allocation of emergency supplies in the warehousing network based on the location decision and pre-reservation decision results in Phase 1. It introduces intermediate variables to transform the Phase 1 model into a scenario probability constraint model and performs dual transformation.

2. The warehouse network model based on two-stage model collaborative optimization according to claim 1, characterized in that, The analysis and prediction module includes: The historical data analysis unit is used to collect past order data and market sales data, and to use time series analysis to analyze the changing trends of demand over time, seasonal fluctuation patterns, and their correlation with other variables. The forecasting unit is used to dynamically forecast demand by combining demand trends over time, seasonal fluctuations, correlations with other variables, industry dynamics, and changes in consumer preferences. Based on the results of the dynamic demand forecast, the storage capacity of each node in the warehouse is adjusted.

3. The warehouse network model based on two-stage model collaborative optimization according to claim 1, characterized in that, The inventory collaborative management module includes: The site selection decision unit is used to determine the warehouse area and floor height parameters based on the storage capacity of each node in the warehouse, and to obtain key constraints and target priorities from five dimensions to determine the warehouse site selection decision. The pre-reservation decision unit is used to clarify the decision boundary and input conditions. Combining the two-stage collaborative logic, it performs calculations according to the steps of planning rough calculation, operation fine calculation and collaborative optimization. It is also classified and stratified according to cargo type, node function and scenario requirements to obtain the pre-reservation decision. The calculation unit is used to calculate the objective function, material supply warehouse location constraints, and warehouse capacity constraints for Phase 1, as well as the network node flow balance constraints, network arc capacity constraints, and decision variable constraints for Phase 2.

4. The warehouse network model based on two-stage model collaborative optimization according to claim 3, characterized in that, The pre-reservation decision unit specifically includes: Obtain key information across five dimensions: demand, supply, network planning, cost, and external risks, and clarify decision-making boundaries and input conditions; Based on the decision boundary and input conditions, and combined with the two-stage collaborative logic, the calculation is performed according to the steps of rough planning calculation, operational fine calculation and collaborative optimization to obtain the specific value of the pre-reserve. Based on the calculation results, the cargo type, node function, and scenario requirements are combined to classify and stratify the cargo to obtain the pre-storage decision.

5. A warehouse network model based on two-stage model collaborative optimization according to claim 3, characterized in that, The computing unit specifically includes: The Phase 1 model is used to calculate the objective function, material supply warehouse location constraints, and warehouse capacity constraints for Phase 1. The Phase 2 model is used to calculate the network node flow balance constraints in Phase 2, as well as the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

6. A warehouse network model based on two-stage model collaborative optimization according to claim 3, characterized in that, The Phase 1 model includes: The objective function for Phase 1 is calculated by minimizing the supply warehouse location cost, the emergency material pre-stocking procurement cost, and the expected cost under the worst-case demand distribution in Phase 2 after the location and pre-stocking quantities are determined. The calculation formula is as follows: ； In the formula, For site selection decisions, This is the transportation cost coefficient. For pre-reserve quantity decision, To determine the expected costs resulting from the pre-reservation and site selection decisions under the worst-case demand distribution in Stage 2, This is the inventory holding cost coefficient; Calculate the site selection constraints for material supply warehouses, namely, the number of supply points selected at each node must not exceed 1; Calculate warehouse capacity constraints, namely, the amount of pre-stored materials must not exceed the corresponding capacity of the supply warehouse and decision variable constraints.

7. A warehouse network model based on two-stage model collaborative optimization according to claim 3, characterized in that, The Phase 2 model includes: The network node flow balance constraint in Phase 2 is calculated by subtracting the outflow and demand of each node from the sum of the inflow and pre-stored resources. If the difference is positive, the node will have surplus resources; otherwise, a negative value indicates a shortage of resources. Calculate the network arc capacity constraints and decision variable constraints that must be satisfied during transportation after a sudden event.

8. A warehouse network model based on two-stage model collaborative optimization according to claim 1, characterized in that, The optimization module includes: The allocation unit is used to plan and allocate emergency supplies in the warehousing network based on the site selection and pre-stocking decisions made in Phase 1. The optimization algorithm unit is used to introduce intermediate variables to transform the stage one model into a scenario probability constraint model, transform the scenario probability constraint model into a dual problem, and after the dual transformation, transform the stage two model into a linear optimization model and calculate the dual problem with respect to uncertain scenario probabilities.

9. A warehouse network model based on two-stage model collaborative optimization according to claim 8, characterized in that, The introduction of intermediate variables to transform the Phase 1 model into a scenario probability constraint model specifically includes: By substituting intermediate variables into the objective function of the stage one model, replacing the original fixed parameters, and incorporating the probability of scenario occurrence, the objective function is transformed from minimizing deterministic costs to minimizing expected costs that take into account scenario probabilities. The deterministic constraints of the Phase 1 model are transformed into constraints that change dynamically with the scenario and meet probability requirements through intermediate variables. The transformed scenario probability constraint model is solved using an algorithm, and the rationality of the intermediate variables and the robustness of the model are verified.

10. A two-stage model collaborative optimization algorithm, implemented based on the warehouse network model based on the two-stage model collaborative optimization described in claim 9, characterized in that, Includes the following steps: S1: Determine the optimal geographical locations of factories, transit stations, and warehouse nodes through spatial analysis, perform dynamic demand forecasting, and dynamically adjust the storage capacity of each node based on the forecast results; S2: Obtain key constraints and target priorities, determine warehouse location decisions, and classify them in layers based on cargo type, node function and scenario requirements to obtain pre-storage decisions; S3: Calculate the objective function, material supply warehouse location constraints, warehouse capacity constraints, and decision variable constraints for Phase 1; S4: Calculate the network node flow balance constraints in Phase 2, the network arc capacity constraints to be satisfied during transportation after a sudden event, and the constraints of the calculation decision variables; S5: Based on the site selection and pre-stocking decisions in Phase 1, plan the scheduling and allocation of emergency supplies in the warehousing network, and calculate the dual problem concerning the probability of uncertain scenarios.

Images