Supply chain performance path optimization and collaborative decision-making method based on AI algorithm
Through the supply chain fulfillment path optimization method based on AI algorithm, the problems of decision-making logic fragmentation and risk response lag in supply chain path optimization are solved, single-warehouse and multi-warehouse collaborative decision-making is realized, and resource allocation efficiency and adaptability to complex scenarios are improved.
Patent Information
- Application Number
- CN202511269576.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-08
- Publication Date
- 2025-10-03
- Estimated Expiration
- 2045-09-08
AI Technical Summary
Existing supply chain path optimization technologies have problems such as fragmented decision-making logic, single evaluation dimension and delayed risk response, which lead to high resource mismatch rate, weak resistance to sudden risks and insufficient adaptability to complex scenarios.
A supply chain fulfillment path optimization method based on AI algorithm is adopted. By obtaining order and warehouse information, a comprehensive evaluation index function is constructed. Combined with the transportation cost and risk prediction function, the optimal path and concentration point are dynamically selected to achieve collaborative decision-making for single and multi-warehouses.
It improves the efficiency of cross-warehouse resource allocation, reduces the probability of default, enhances the adaptability and anti-interference ability of the supply chain in complex scenarios, and optimizes resource utilization.
Smart Images

Figure CN120745992A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of supply chain fulfillment path optimization, and more specifically, to a supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm. Background Art
[0002] Current supply chain path optimization technology has developed a data-driven methodology. Mainstream solutions focus on single-warehouse intelligent scheduling, multi-warehouse collaboration, and risk response mechanisms. In single-warehouse scenarios, AI algorithms analyze historical data to generate local optimal paths, significantly improving warehouse efficiency. In multi-warehouse collaboration, centralized systems leverage digital twin technology to achieve a preliminary match between capacity and tasks. Regarding risk prevention and control, leading companies integrate data from multiple sources, such as meteorology and transportation, to build predictive models, enhancing their ability to anticipate delays. Emerging technologies, such as big models, are beginning to penetrate logistics decision-making systems, driving the evolution of supply chains towards dynamism and visualization. However, these technologies remain limited by centralized architectures, with optimization focused on localized links. Global resource coordination and real-time risk management capabilities have yet to be achieved.
[0003] However, it still has some shortcomings in actual use: Fragmented decision-making logic: Single-warehouse optimization and multi-warehouse collaboration are handled in isolation. Most systems are unable to dynamically switch decision-making modes based on order demand, resulting in inefficient cross-warehouse resource allocation. Single evaluation dimension: Route optimization relies too much on transportation time and distance cost, ignoring key implicit indicators such as route complexity and time buffer redundancy, resulting in significant deviations in actual execution; Delayed risk response: Traditional risk models rely on static historical data and are unable to dynamically integrate sudden environmental disturbances. Risk prediction is disconnected from cost optimization, and emergency decision-making relies on human intervention.
[0004] These defects together lead to systemic bottlenecks in the supply chain, such as high resource mismatch rate, weak resistance to sudden risks, and insufficient adaptability to complex scenarios. Summary of the Invention
[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm, which solves the problems raised in the above-mentioned background technology through the following scheme.
[0006] To achieve the above objectives, the present invention provides the following technical solution: a supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm, comprising: S1: Data acquisition: Obtain target order information and warehouse site information for pre-processing; S2: Feasibility assessment of single-warehouse fulfillment: Match the target order information with the warehouse site information to determine whether a single warehouse can meet all the requirements for the order goods. If so, determine the candidate warehouse and proceed to step S3; otherwise, proceed to step S4. S3: Generate the optimal path for a single warehouse: Use the candidate warehouse as the starting point and the delivery address in the order information as the end point, and construct a comprehensive evaluation index function. Use this comprehensive evaluation index function to evaluate each path from the starting point to the end point and determine the optimal path. If there are two or more candidate warehouses, determine the optimal candidate warehouse based on the comprehensive evaluation index of each warehouse's optimal path. S4: Multi-warehouse collaborative decision-making and route optimization: When multiple warehouses need to coordinate to meet all the demand for target order goods, these warehouses are selected as candidate warehouses for the centralized point, and the transfer warehouse is determined. Then, a function is established and analyzed to obtain the multi-warehouse coordinated fulfillment path. The specific steps are as follows: S401: Constructing a transshipment cost function: Calculating the total transportation cost and total transportation time of the goods required to be transshipped from other warehouses to each candidate warehouse as the central point; S402: Construct a risk prediction function: Based on the warehouse's historical performance time and external environmental risks, calculate the default risk value of the warehouse as a concentration point; S403: Based on the transshipment cost function and the default risk prediction function, a comprehensive early warning function is generated for each candidate warehouse through weighted calculation, and the candidate warehouse with the best comprehensive evaluation value is selected as the final concentration point; S404: Based on the determined concentration point, plan the transfer path from each warehouse to the concentration point and the fulfillment path from the concentration point to the order delivery address to form a multi-warehouse collaborative fulfillment path.
[0007] Preferably, the data acquisition includes acquiring target order information and information of each warehouse site and performing pre-processing; The target order information obtained includes: type of goods, quantity of goods, delivery deadline, and delivery address; the warehouse site information includes: inventory data and transportation resource data of each warehouse; Among them, inventory data includes: inventory types, inventory quantities, and unit volumes of each type of goods; transportation resource data includes: vehicle resources, vehicle capacity, vehicle load, and loading and unloading efficiency; and order volume, order weight, and estimated loading and unloading time are calculated based on target order information and inventory data.
[0008] Preferably, the feasibility assessment of the single warehouse performance includes: Based on the goods types and quantities in the target order information, the inventory goods quantities and types in the warehouse site information are matched. If there is a single warehouse that meets all the conditions of the target order information, then all the one or more single warehouses that meet the conditions are taken as candidate warehouses and S3 is executed; if there is no one or more single warehouses that meet the conditions, but the total inventory goods quantity and types of multiple single warehouses meet all the matching conditions of the target order information, then each single warehouse that meets the conditions is taken as a candidate warehouse for the centralized point and enters S4.
[0009] Preferably, the generation of the single-warehouse optimal path includes: Taking the candidate warehouse as the starting point and the delivery address in the order information as the end point, we count all possible path solutions from the starting point to the end point and predict the estimated travel time of all possible paths. Combining order information and transportation resource data, we obtain the estimated delivery time and the set of all available transportation resources. We then generate an evaluation vector for each candidate route that maximizes four metrics: total transportation time, time slack, path complexity, and resource utilization. We then normalize the evaluation metrics of all candidate route plans and calculate a comprehensive evaluation metric. The specific steps are as follows: S301 defines a multi-objective vector: constructs a four-dimensional evaluation vector for each candidate path solution, including total transportation time, time redundancy, path complexity, and resource utilization; S302 builds a standardized decision matrix: generates an m*4 dimensional standard matrix based on vector normalization method; S303 Determine ideal values and anti-ideal values: Calculate the maximum value of each indicator in the solution to form an ideal solution, and calculate the minimum value to form a negative ideal solution; S304 Calculate distance measure: introduce weight coefficients, weighted processing to strengthen the impact of key indicators, and quantify the degree of deviation between the solution and the ideal solution and the negative ideal solution; S305 calculates a comprehensive evaluation index for a single warehouse: The ratio of the anti-ideal distance to the total distance is used. If a time limit or buffer threshold is violated, the solution is directly eliminated. The scores of all solutions are mapped to the interval [0, 1]. Based on the comprehensive evaluation index value, the optimal path solution is obtained. S306: If there are multiple candidate warehouses, calculate the comprehensive evaluation index of the optimal path for each warehouse, and determine the optimal candidate warehouse based on the comprehensive evaluation index value; Preferably, constructing the transport cost function includes: First, the transshipment cost is calculated for each warehouse that needs to transship to the designated centralized warehouse. Then, the cost of all individual transshipment warehouses is summed up to obtain the transshipment cost function. The cost of an individual transshipment warehouse consists of three parts: distance cost, time cost, and cargo characteristics. A distance cost function is constructed based on the transport distance parameter. This function combines the basic linear coefficient and the nonlinear penalty coefficient to obtain the adjusted basic distance cost and the converged long-distance implicit cost. The tolls generated during the passage are then calculated and combined to obtain the distance cost function. A time cost function is constructed based on the timeliness parameter. This function combines the basic time cost, the emergency transportation coefficient, and the time sensitivity coefficient to obtain the basic time cost and the cargo characteristic cost adjusted according to the characteristics of the cargo itself. The combination of the basic time cost and the cargo characteristic cost is the time cost function. By calculating the cargo density factor and the transshipment cargo volume, the above two types of functions and values are adjusted to obtain the comprehensive transshipment cost function affected by the cargo characteristics.
[0010] Preferably, the constructing of the risk prediction function includes: Risks are quantified from two dimensions: fulfillment timeliness risk and external environment risk. Fulfillment timeliness risk is based on the historical transportation task data of candidate centralized warehouses. It analyzes the deviation between the actual time and the planned time for completing transportation tasks in the past. The deviation rate of all historical tasks of the warehouse is statistically analyzed and combined with the task weights to calculate the weighted average deviation value. The higher the deviation value, the greater the probability of delivery delays when the warehouse is used as a centralized point. External environment risk dynamically integrates the impact of sudden external factors on the transportation route, including implementing meteorological and geographical location sensitivity, and increasing the hidden risks brought by the geographical location of border or port warehouses. The two types of risk values are normalized to the same dimension. The weighted average of historical deviations is directly mapped to a probability value. The external environmental risk is suppressed by the attenuation coefficient to suppress the excessive dominance of external risks. Then, combined with the geographical location sensitivity coefficient adjustment, it is converted into a standardized score. Assign weights to the two types of risks based on business needs, and take their weighted sum to generate a comprehensive risk quantification value.
[0011] Preferably, the comprehensive evaluation function of each candidate warehouse includes: The transshipment cost and risk prediction value of each candidate warehouse are standardized and mapped to the interval [0, 1]. Weights are assigned according to their importance in decision-making. The normalized transshipment cost and risk prediction value are weighted summed. Finally, the comprehensive warning values of all candidate concentration point warehouses are ranked, and the warehouse with the lowest warning value is selected as the final concentration point.
[0012] Preferably, the multi-warehouse collaborative fulfillment path includes: With the determined centralized warehouse as the starting point and the order delivery address as the end point, the optimal fulfillment path from the centralized warehouse to the delivery address is generated by calling the single warehouse optimal path generation method; Taking each transfer warehouse as the starting point and the centralized warehouse as the end point, the optimal transfer path from each transfer warehouse to the centralized point is generated by calling the single-warehouse optimal path generation method; The generated optimal transfer path is combined with the optimal fulfillment path to form a complete multi-warehouse collaborative fulfillment path of transfer warehouse-centralized warehouse-order delivery address.
[0013] Technical effects and advantages of the present invention: 1. Innovative dual-mode dynamic decision-making architecture: This architecture breaks through the limitations of independent traditional decision-making logic and enables adaptive switching of decision-making modes through single-warehouse fulfillment feasibility assessment. In single-warehouse scenarios, a four-dimensional evaluation system is constructed, encompassing total transportation time, time redundancy, path complexity, and resource utilization. Through standardization and distance measurement, the quality of paths is quantified to generate a comprehensive optimal path. In multi-warehouse scenarios, a dynamic centralized point selection model is used to generate a transshipment network based on the distance and inventory matching rules of transshipment warehouses, enabling global resource scheduling and improving cross-warehouse collaboration efficiency.
[0014] 2. Establish a joint risk and cost optimization mechanism. The transshipment cost function integrates distance and time costs, including a basic distance coefficient, long-distance costs, additional costs at toll points, basic time costs, an urgency coefficient, and a sensitivity coefficient, accurately quantifying the differences in transshipment of different goods. The risk prediction function integrates historical performance time deviations and dynamic external environmental risks to generate a normalized risk value. These two factors are weighted and integrated through a comprehensive early warning score to achieve a balance between cost and risk, reducing the probability of default and ineffective transshipment.
[0015] 3. Explicitly quantify implicit factors to enhance adaptability to complex scenarios. Incorporating traditionally overlooked factors such as path complexity and time redundancy into the decision-making system will improve adaptability to complex road networks and urgent orders, and optimize resource utilization.
[0016] 4. A global real-time response mechanism improves system agility. Multiple warehouse scenarios reuse single-warehouse path optimization algorithms to reduce repeated calculations. Dynamically update parameters and adjust paths when the external environment changes, aligning local decisions with global goals and enhancing the supply chain's resilience and operational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 It is a schematic diagram of the overall structure of the present invention.
[0018] Figure 2 It is a schematic diagram of the overall process of the present invention.
[0019] Figure 3 This is a schematic diagram of the single warehouse fulfillment structure of the present invention.
[0020] Figure 4 This is a schematic diagram of the multi-warehouse collaborative decision-making structure of the present invention. DETAILED DESCRIPTION
[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0022] refer to Figure 1 - Figure 4 The following is a supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm, including: S1: Data acquisition: Obtain target order information and warehouse site information for pre-processing; Through the deployment of standardized API interface system and AI algorithm, data automatic processing is achieved: Obtain target order information through the system API interface, automatically capture key order fields, and read the warehouse site information in the local database; The target order information obtained includes: type of goods, quantity of goods, delivery deadline, and delivery address; The information of each warehouse site includes: inventory data and transportation resource data of each warehouse; Among them, inventory data includes: inventory type, inventory quantity, unit volume of each cargo, and unit weight of each cargo; transportation resource data includes: vehicle resources, vehicle capacity, vehicle load, and loading and unloading efficiency; The unit volume parameters of each cargo in the local database are called, and the order cargo volume and order cargo weight are calculated based on the cargo type, cargo quantity, and unit weight of each cargo in the order. The estimated loading and unloading time is calculated based on the loading and unloading efficiency.
[0023] S2: Feasibility determination of single-warehouse fulfillment: Match the target order information with the warehouse site information to determine whether there is a single warehouse that meets all the requirements of the order goods, and determine the candidate warehouses; if there is a single warehouse that meets all the requirements of the order goods, execute step S3, otherwise execute step S4.
[0024] Based on the target order information's goods type and quantity, the inventory quantity and types of goods in each warehouse site information are matched to determine whether there is a single warehouse that meets the target order information's goods type and quantity requirements. If there is a single warehouse that meets all the conditions of the target order information, then all of the one or more single warehouses that meet the conditions are taken as candidate warehouses, and step S3 is executed; If there are not one or more single warehouses that meet the conditions, but the total quantity and types of inventory goods of multiple single warehouses meet all matching conditions of the target order information, each single warehouse that meets the conditions will be used as a candidate warehouse for the concentration point and enter S4.
[0025] S3: Single-warehouse optimal path generation: Taking the candidate warehouse as the starting point and the delivery address in the order information as the end point, a graph neural network model is constructed based on the AI algorithm. The set of all feasible path solutions from the starting point to the end point is counted and recorded as: ,in Denotes the i-th path plan, predicting the estimated travel time of all accessible paths; and based on the transportation resource data in the order information and warehouse site information, gives the set of all available transportation resources, which is recorded as: , where h represents the total number of all transportation resources; based on this, a comprehensive evaluation index of the path is generated and the optimal path is selected. The specific steps are as follows: S301: Define a multi-objective vector: Construct a four-dimensional evaluation vector for each candidate path solution, unifying the four key optimization objectives into a maximization problem: ; in: It represents the negative value of the total transportation time, and T represents the sum of the estimated travel time and the estimated loading and unloading time; It represents the time redundancy, that is, the margin that the arrival time is earlier than the order delivery deadline. ,in, Indicates the order delivery deadline, represents the estimated delivery arrival time of plan i; Indicates the negative value of path complexity, , where edges represents the road section type, including: expressway, main road, congested road section, construction road section, and its corresponding complexity coefficient are: 0.1, 0.3, 0.8, 1.5; Indicates resource utilization, ; S302: Building a standardized decision matrix Define the initial decision matrix: combine the evaluation vectors of m candidate solutions into an m*4 matrix: ; Where m is the total number of candidate solutions, represents the jth indicator value of the i-th solution. The first column represents the negative value of the transportation time, the second column represents the time redundancy, the third column represents the negative value of the path complexity, and the fourth column represents the resource utilization rate. Then, normalization is performed to eliminate the unit difference of the indicators and avoid large-scale indicators dominating the calculation results. The matrix Z is obtained. The specific mathematical function of its normalization is: , in Represents the standardized value of the i-th scheme and the j-th indicator; the standardized matrix is obtained: ; S303: Determine ideal value and anti-ideal value Based on the standardized decision matrix Z, the theoretically optimal solution and the worst solution are defined. The ideal solution serves as the benchmark for subsequent distance calculations. The ideal solution is a virtual optimal solution that takes the maximum value of all candidate solutions in each indicator. The calculation formula is: ; in represents the maximum value of the jth indicator among the m solutions; Anti-ideal solution: A virtual worst-case solution that takes the minimum value of all candidate solutions on each indicator. Its calculation formula is: ; in represents the minimum value of the jth indicator among the m solutions; S304: Calculate distance measure Calculate each candidate solution i to the ideal solution and anti-ideal solutions The weighted Euclidean distance of quantifies the deviation between the solution and the optimal and worst solutions; the distance between solution i and the ideal solution The specific mathematical function of the weighted Euclidean distance is: ; Solution i to ideal solution The specific mathematical function of the weighted Euclidean distance is: ; in represents the standardized value of the jth indicator for the i-th scheme, is the weight of the jth indicator. According to the actual situation, the priority ranking should be based on efficiency, reliability, road complexity, and cost. , whose constraints are: ; S305: Calculate the comprehensive evaluation index of a single warehouse By calculating the comprehensive evaluation index of each scheme i , comprehensively evaluate its closeness to the ideal solution and its distance from the anti-ideal solution, and generate a single-bin comprehensive evaluation index in the interval [0, 1]. Its specific mathematical function is: ; It should be further explained that all accessible routes must meet the following conditions when they are determined to be the optimal route: 1. The estimated transportation time of the route must not exceed the order delivery deadline; 2. The redundant time must not be less than the preset minimum buffer time; According to the above steps, the optimal path solution is obtained: , indicating the selection of comprehensive evaluation The highest solution P is taken as the optimal path solution under the transportation resources; S306: Determine the optimal candidate warehouse If there are two or more candidate warehouses, repeat the above steps to calculate the optimal path plan for each candidate warehouse, and sort them from large to small according to the comprehensive evaluation indicators under their optimal path plans. Select the candidate warehouse with the largest comprehensive evaluation indicator and determine it as the optimal warehouse.
[0026] S4: Multi-warehouse collaborative decision-making and route optimization: Each warehouse that requires coordinated deployment of multiple warehouses is selected as a candidate warehouse for the centralized point. The transshipment cost function and transshipment risk function of each candidate warehouse for the centralized point are constructed to obtain a comprehensive feasibility function. The optimal centralized point warehouse is determined based on the function value. The specific steps are as follows: The candidate centralized warehouse is defined as y, and the transfer warehouse set is defined as , Represents the xth transshipment warehouse. The definition principle of the transshipment warehouse set is: based on the quantity of goods missing from the candidate centralized warehouse and the inventory of goods in all the transshipment warehouses, goods are allocated from near to far according to the distance between the warehouses, that is, first sort the distances between all transshipment warehouses and the candidate centralized warehouses from short to long. If the inventory of the candidate transshipment warehouse with the first ranking result meets the quantity of goods missing from the candidate centralized warehouse, then only this candidate transshipment warehouse will be used as the transshipment warehouse; if not, then determine whether the total inventory of the first and second transshipment warehouses with the ranking results meets the quantity of goods missing from the candidate centralized warehouse. If so, then the first and second candidate transshipment warehouses with the ranking results will be used as transshipment warehouses; if not, then calculate whether the total inventory of the first, second, and third transshipment warehouses with the ranking results meets the requirements, until the total inventory of the candidate transshipment warehouse meets the quantity of goods missing from the candidate centralized warehouse, then the candidate transshipment warehouse that meets the requirements will be used as the transshipment warehouse for the candidate centralized warehouse. It should be further explained that here the candidate concentration point is defined as the end point, the transfer warehouse is defined as the starting point, and steps S301-S305 are repeated to obtain the optimal path plan Pi from the transfer warehouse to the candidate concentration point warehouse, and the actual distance from the transfer warehouse x to the candidate concentration point y is obtained based on the optimal path , estimated time , and calculate the proportion of perishable goods in the transshipment cargo volume; S401: Constructing the transshipment cost function: The transshipment cost is defined as , its specific mathematical function is as follows:
[0027] in is the transshipment volume of warehouse x, is the cargo density factor, ;Distance cost function: ,in is the basic linear coefficient, which represents the basic operating cost per unit distance and reflects the fuel consumption and vehicle loss per unit distance; is a nonlinear penalty coefficient, which represents the marginal cost increasing factor of long-distance transportation and reflects the implicit cost, but it does not dominate the distance cost, so it is set to 0.02. is the distance index, with a value of 1.3. is the number of charging points, It is a single toll fee; Time cost function: ,in As the basic time cost, For the emergency coefficient, when transporting ordinary goods, , when transporting fragile goods, , is the time sensitivity, with a value of 0.4; S402: Construct a risk prediction function: Quantify the risk of order fulfillment failure due to transportation delays or external unexpected factors when the candidate centralized warehouse y is used as the centralized point, and output a normalized score R. The mathematical function is as follows:
[0028] in, is the time risk weight, The environmental risk weight is 0.4, and the environmental risk weight is 0.6. Expressed as the fulfillment time risk, based on the historical transportation task time deviation of warehouse y, the probability of delivery delay when it is the concentration point is predicted, which is expressed as:
[0029] in, is the sample size of historical transportation tasks for warehouse y, is the actual transportation time of task j, is the planned transportation time for task j, is the task weight when transporting fragile goods , when transporting general cargo , the historical task volume is the historical completed tasks within 90 days; It is expressed as external environmental risk, dynamically integrating the impact of sudden factors on the transportation path, which is expressed as: ,in is the attenuation coefficient to avoid external risks dominating the total score, , is the geographical location sensitivity coefficient, ordinary warehouse: , border or port warehouse , is the safety threshold, the value is 2, Implement risk index for the region, dynamically integrate real-time meteorological factors based on access to the meteorological platform API, and update real-time safety thresholds , the value is [0, 4].
[0030] S403: Constructing a comprehensive warning function for candidate concentration points: By weighted fusion of the transshipment cost function and the risk prediction function, a comprehensive warning function for candidate concentration points is constructed. The design logic is as follows: the transshipment cost and risk prediction value of each candidate warehouse are normalized and mapped to the interval [0, 1]. Weights are then assigned based on the importance of the transshipment cost and risk prediction value in the decision-making process. The normalized transshipment cost and risk prediction value are then multiplied by the corresponding weights and added together to obtain a comprehensive warning value for each candidate concentration point warehouse. Finally, the comprehensive warning values of all candidate concentration point warehouses are sorted, and the warehouse with the lowest warning value is selected as the final concentration point. The specific steps are as follows: Normalization processing: , The combined early warning function: ,in 、 are the weight coefficients of transshipment cost and risk prediction value respectively. Since the warehouse closest to the candidate centralized warehouse is given priority during transshipment, the transshipment cost can more prominently reflect the comprehensive warning value of the candidate centralized warehouse. , ; Calculate the comprehensive warning values of all candidate centralized warehouses according to the above steps, and sort them from low to high according to their values. Select the candidate centralized warehouse with the lowest comprehensive warning value as the centralized warehouse.
[0031] S404: Based on the determined centralized warehouse, plan the transfer routes from each transfer warehouse to the centralized warehouse and the fulfillment route from the centralized warehouse to the order delivery address to form a multi-warehouse collaborative fulfillment route.
[0032] Using the determined centralized warehouse as the starting point and the delivery address in the target order information as the end point, repeat step S3 to determine the optimal path from the centralized warehouse to the order delivery address; Take the centralized warehouse as the end point, and then take the warehouses that need to transfer the centralized warehouse as the starting points, and repeat step S3 to determine the optimal path from the transfer warehouse to the centralized warehouse. Based on the optimal path from the transshipment warehouse to the centralized warehouse and the optimal path from the centralized warehouse to the order delivery address, a multi-warehouse collaborative optimal fulfillment path is formed from the transshipment warehouse to the centralized warehouse and finally to the order delivery address.
[0033] This invention systematically solves the core defects of supply chain path optimization, such as fragmented resource scheduling, delayed risk response, and single evaluation dimension, by building a dual-path decision-making framework for single-warehouse fulfillment and multi-warehouse collaboration. Its innovative implementation is reflected in three levels: Basic layer: Dynamically judge the order fulfillment model based on AI algorithms: the feasibility of a single warehouse and the necessity of multi-warehouse collaboration, and realize intelligent switching of resource scheduling modes; Optimization layer: In single-warehouse scenarios, a multi-dimensional path evaluation system is used to generate a local optimal path based on transportation efficiency, time redundancy, path complexity, and resource utilization. An innovative comprehensive early warning scoring model is designed for multi-warehouse scenarios. By weighted integration of the transshipment cost function and the risk prediction function, the global optimal concentration point is dynamically selected. Execution layer: Reuse the single-warehouse path optimization algorithm to generate a full-link collaborative solution from the transfer warehouse to the centralized point and finally to the end user, achieving seamless connection between resource allocation and fulfillment path.
[0034] This method fundamentally breaks through the limitations of static planning and empirical decision-making in traditional technologies, drives global resource optimization with dynamic risk prediction, significantly improves the adaptability, robustness and execution efficiency of the supply chain in complex scenarios, and provides a scalable intelligent decision-making basis for highly volatile logistics environments.
[0035] Secondly: The drawings of the embodiments disclosed in the present invention only involve structures related to the embodiments disclosed in the present invention. Other structures may refer to conventional designs. The same embodiment and different embodiments of the present invention may be combined with each other without conflict. Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm, characterized by: include: S1: Data acquisition: Obtain target order information and warehouse site information for pre-processing; S2: Feasibility assessment of single-warehouse fulfillment: Match the target order information with the warehouse site information to determine whether a single warehouse can meet all the requirements for the order goods; If so, determine the candidate warehouse and execute step S3; otherwise, execute step S4; S3: Generate the optimal path for a single warehouse: Use the candidate warehouse as the starting point and the delivery address in the order information as the end point, and construct a comprehensive evaluation index function. Use this comprehensive evaluation index function to evaluate each path from the starting point to the end point and determine the optimal path. If there are two or more candidate warehouses, determine the optimal candidate warehouse based on the comprehensive evaluation index of each warehouse's optimal path. S4: Multi-warehouse collaborative decision-making and route optimization: When multiple warehouses need to coordinate to meet all the demand for the target order goods, these warehouses are selected as candidate centralized warehouses, and the transshipment warehouses are determined. Then, a function is established and analyzed to obtain the multi-warehouse coordinated fulfillment path. The specific steps are as follows: S401: Constructing a transshipment cost function: Calculating the total transportation cost and total transportation time of the goods required to be transshipped from other warehouses to each candidate warehouse as the central point; S402: Construct a risk prediction function: Based on the warehouse's historical performance time and external environmental risks, calculate the default risk value of the warehouse as a concentration point; S403: Based on the transshipment cost function and the default risk prediction function, a comprehensive early warning function is generated for each candidate warehouse through weighted calculation, and the candidate warehouse with the best comprehensive evaluation value is selected as the final concentration point; S404: Based on the determined concentration point, plan the transfer path from each warehouse to the concentration point and the fulfillment path from the concentration point to the order delivery address to form a multi-warehouse collaborative fulfillment path.
2. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The data acquisition includes acquiring target order information and warehouse site information and performing pre-processing; The target order information obtained includes: type of goods, quantity of goods, delivery deadline, and delivery address; the warehouse site information includes: inventory data and transportation resource data of each warehouse; Among them, inventory data includes: inventory types, inventory quantities, and unit volumes of each type of goods; transportation resource data includes: vehicle resources, vehicle capacity, vehicle load, and loading and unloading efficiency; and order volume, order weight, and estimated loading and unloading time are calculated based on target order information and inventory data.
3. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The feasibility assessment of the single warehouse performance includes: Based on the goods types and quantities in the target order information, the inventory goods quantities and inventory goods types in the warehouse site information are matched. If there is a single warehouse that meets all the conditions of the target order information, then all the one or more single warehouses that meet the conditions are taken as candidate warehouses and S3 is executed; if there is no one or more single warehouses that meet the conditions, but the total inventory goods quantity and types of multiple single warehouses meet all the matching conditions of the target order information, then each single warehouse that meets the conditions is taken as a candidate warehouse for the centralized point and enters S4.
4. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The generation of the single-warehouse optimal path includes: Taking the candidate warehouse as the starting point and the delivery address in the order information as the end point, we count all possible path solutions from the starting point to the end point and predict the estimated travel time of all possible paths. Combining order information and transportation resource data, we obtain the estimated delivery time and the set of all available transportation resources. We then generate an evaluation vector for each candidate route that maximizes four metrics: total transportation time, time slack, path complexity, and resource utilization. We then normalize the evaluation metrics of all candidate route plans and calculate a comprehensive evaluation metric. The specific steps are as follows: S301 defines a multi-objective vector: constructs a four-dimensional evaluation vector for each candidate path solution, including total transportation time, time redundancy, path complexity, and resource utilization; S302 builds a standardized decision matrix: generates an m*4 dimensional standard matrix based on vector normalization method; S303 Determine ideal values and anti-ideal values: Calculate the maximum value of each indicator in the solution to form an ideal solution, and calculate the minimum value to form a negative ideal solution; S304 Calculate distance measure: introduce weight coefficients, weighted processing to strengthen the impact of key indicators, and quantify the degree of deviation between the solution and the ideal solution and the negative ideal solution; S305 calculates a comprehensive evaluation index for a single warehouse: The ratio of the anti-ideal distance to the total distance is used. If a time limit or buffer threshold is violated, the solution is directly eliminated. The scores of all solutions are mapped to the interval [0, 1]. Based on the comprehensive evaluation index value, the optimal path solution is obtained. S306: If there are multiple candidate warehouses, calculate the comprehensive evaluation index of the optimal path for each warehouse, and determine the optimal candidate warehouse based on the comprehensive evaluation index value.
5. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The constructing of the transport cost function includes: First, the transshipment cost is calculated for each warehouse that needs to transship to the designated centralized warehouse. Then, the cost of all individual transshipment warehouses is summed up to obtain the transshipment cost function. The cost of an individual transshipment warehouse consists of three parts: distance cost, time cost, and cargo characteristics. A distance cost function is constructed based on the transport distance parameter. This function combines the basic linear coefficient and the nonlinear penalty coefficient to obtain the adjusted basic distance cost and the converged long-distance implicit cost. The tolls generated during the passage are then calculated and combined to obtain the distance cost function. A time cost function is constructed based on the timeliness parameter. This function combines the basic time cost, the emergency transportation coefficient, and the time sensitivity coefficient to obtain the basic time cost and the cargo characteristic cost adjusted according to the characteristics of the cargo itself. The combination of the basic time cost and the cargo characteristic cost is the time cost function. By calculating the cargo density factor and the transshipment cargo volume, the above two types of functions and values are adjusted to obtain the comprehensive transshipment cost function affected by the cargo characteristics.
6. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The constructing of the risk prediction function includes: Risks are quantified from two dimensions: fulfillment timeliness risk and external environment risk. Fulfillment timeliness risk is based on the historical transportation task data of candidate centralized warehouses. It analyzes the deviation between the actual time and the planned time for completing transportation tasks in the past. The deviation rate of all historical tasks of the warehouse is statistically analyzed and combined with the task weights to calculate the weighted average deviation value. The higher the deviation value, the greater the probability of delivery delays when the warehouse is used as a centralized point. External environment risk dynamically integrates the impact of sudden external factors on the transportation route, including implementing meteorological and geographical location sensitivity, and increasing the hidden risks brought by the geographical location of border or port warehouses. The two types of risk values are normalized to the same dimension. The weighted average of historical deviations is directly mapped to a probability value. The external environmental risk is suppressed by the attenuation coefficient to suppress the excessive dominance of external risks. Then, combined with the geographical location sensitivity coefficient adjustment, it is converted into a standardized score. Assign weights to the two types of risks based on business needs, and take their weighted sum to generate a comprehensive risk quantification value.
7. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The comprehensive early warning function of each candidate warehouse includes: The transshipment cost and risk prediction value of each candidate warehouse are standardized and mapped to the interval [0, 1]. Then, weights are assigned according to their importance in decision-making. The normalized transshipment cost and risk prediction value are weighted and summed. Finally, the comprehensive warning values of all candidate concentration point warehouses are ranked from low to high, and the warehouse with the lowest warning value is selected as the final concentration point.
8. The supply chain fulfillment path optimization and collaborative decision-making method based on AI algorithm according to claim 1 is characterized by: The multi-warehouse collaborative fulfillment path includes: With the determined centralized warehouse as the starting point and the order delivery address as the end point, the optimal fulfillment path from the centralized warehouse to the delivery address is generated by calling the single warehouse optimal path generation method; Taking each transfer warehouse as the starting point and the centralized warehouse as the end point, the optimal transfer path from each transfer warehouse to the centralized point is generated by calling the single-warehouse optimal path generation method; The generated optimal transfer path is combined with the optimal fulfillment path to form a complete multi-warehouse collaborative fulfillment path from the transfer warehouse to the centralized warehouse and finally to the order delivery address.
Citation Information
Patent Citations
Intelligent multi-warehouse delivery method and system for e-commerce service
CN115760305A
Order source searching algorithm and system for DTC mode multi-warehouse delivery in retail industry
CN116228372A
Data management method and system for supply chain
CN118115000A
Inventory allocation optimization method and system based on genetic algorithm
CN119090408A
Self-adaptive logistics task allocation method for multi-warehouse cooperation
CN120069473A
Cited By
Logistics path optimization method and system, readable storage medium and computer
CN121328881A
Unitized configuration-based spot open management system and method
CN121436829A
Medical supply chain multi-level collaborative intelligent order scheduling and tracking method and system
CN121882862A
A multi-level collaborative intelligent order scheduling and tracking method and system for a medical supply chain
CN121882862B
