Medical consumable supply chain intelligent scheduling method

Abstract

The invention relates to the technical field of medical consumable supply chain scheduling, and discloses a medical consumable supply chain intelligent scheduling method. The method comprises the following steps: firstly, acquiring real-time orders and inventory data of medical consumables, and generating a demand prediction graph and an inventory state graph after standardized preprocessing; constructing a multi-objective optimization scheduling model of a double-layer decision network based on linear programming and dynamic programming, and initializing model parameters; and inputting the preprocessed data into the model, obtaining a scheduling path planning graph and an inventory distribution graph through forward calculation, then calculating deviation loss and matching loss, updating model parameters through back propagation, and outputting a scheduling scheme and an inventory dynamic adjustment strategy after multiple iterative optimization. According to the method, multiple targets such as transportation time and cost can be comprehensively considered, and through real-time feedback correction and multi-dimensional evaluation, the scheduling accuracy and the supply chain response speed are improved, inventory management is optimized, and the overall benefits of the medical consumable supply chain are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical consumables supply chain scheduling, and specifically to an intelligent scheduling method for a medical consumables supply chain. Background Art

[0002] In the rapidly developing medical industry, the efficient supply of medical consumables plays a vital role in ensuring the quality of medical services and patient safety. However, the current supply chain scheduling of medical consumables faces many severe challenges, which seriously restrict the development of the industry. From a data perspective, medical consumables order and inventory data are highly complex and dynamic. Demand for medical consumables varies greatly across different medical institutions and departments, and this demand fluctuates frequently due to factors such as seasonal changes and public health emergencies. At the same time, inventory data is difficult to accurately reflect the actual inventory status in real time due to issues such as irregular inbound and outbound operations and delayed data entry. Traditional data processing methods often remain at the simple recording and statistics stage, unable to deeply mine and effectively integrate these complex data. As a result, the value of the data is not fully utilized, making it difficult to support scientific and reasonable scheduling decisions. In terms of scheduling models, most existing scheduling methods adopt single-objective optimization strategies, such as simply pursuing the lowest transportation cost or the shortest delivery time, while ignoring other key factors. This one-sided optimization approach makes it difficult for scheduling solutions to balance the interests of all parties in actual applications and cannot meet the needs of multi-objective collaborative optimization of the medical consumables supply chain. For example, choosing a longer delivery route to reduce transportation costs may result in medical consumables not being delivered in a timely manner, affecting medical treatment; excessive pursuit of delivery speed may significantly increase transportation costs and reduce the company's economic benefits. In addition, traditional scheduling models lack adaptability to dynamically changing environments, making it difficult to adjust scheduling strategies in real time according to actual conditions. In practice, logistics and distribution processes are plagued by numerous uncertainties, such as traffic congestion and weather changes. These factors can lead to delivery delays and impact the timely supply of medical supplies. Inventory management also presents challenges. Overstocking consumes significant capital and storage space, increasing the risk of loss. Insufficient inventory can hinder the smooth implementation of medical treatment and even endanger patients' lives. Currently, most medical institutions and companies lack effective dynamic inventory adjustment mechanisms, making it difficult to adjust inventory levels in a timely manner based on changes in demand. At the same time, with the continuous development of the medical industry, the requirements for the quality and safety of medical consumables are becoming increasingly stringent. Traditional supply chain scheduling methods are insufficient in ensuring the quality and safety of medical consumables, making it difficult to effectively monitor and manage the transportation and storage conditions of consumables. For example, if the transportation and storage conditions of some medical consumables, which are sensitive to environmental conditions such as temperature and humidity, are not met, their performance and quality may be affected, posing a potential safety hazard to medical care. In summary, the existing medical consumables supply chain scheduling methods have obvious shortcomings in data processing, model optimization, coping with uncertainty, and ensuring the quality and safety of consumables. There is an urgent need for a more intelligent, efficient, and reliable scheduling method to meet the ever-evolving needs of the medical industry. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent scheduling method for the medical consumables supply chain to solve the problems raised in the above background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent scheduling of a medical consumables supply chain, the method comprising: Step S1: Acquire real-time order data and inventory data of medical consumables, perform standardization preprocessing on the data, and generate a demand forecast graph and an inventory status graph; Step S2: constructing a multi-objective optimization scheduling model, wherein the model includes a two-layer decision network based on linear programming and dynamic programming; Step S3: Initialize the parameters of the multi-objective optimization scheduling model, including order priority weights and inventory turnover coefficients; Step S4: input the order data, demand forecast graph, and inventory status graph pre-processed in step S1 into the multi-objective optimization scheduling model, and generate a scheduling path planning graph and an inventory allocation graph through forward calculation; Step S5: Calculate the deviation loss between the scheduling path planning map generated in step S4 and the actual logistics path, calculate the matching loss between the inventory allocation map and the inventory status map, and backpropagate to update the model parameters; Step S6, repeating steps S4 and S5 to a preset number of iterations; Step S7: Based on the optimized multi-objective optimization scheduling model, output the scheduling plan and inventory dynamic adjustment strategy for medical consumables.

[0005] Preferably, the multi-objective optimization scheduling model is based on mixed integer programming. The first layer of the multi-objective optimization scheduling model generates an initial scheduling path through dynamic programming, and the second layer performs multi-stage optimization on the initial path through linear programming. The multi-stage optimization module is composed of two-stage path correction modules connected in series. The path correction module inputs the input path data into the time window constraint analysis layer and the cost constraint analysis layer respectively, and outputs the weighted fusion of the analysis results.

[0006] Preferably, the multi-objective optimization scheduling model performs path smoothing, resource allocation normalization and threshold activation in sequence after the multi-stage optimization module to generate a final scheduling path planning diagram.

[0007] Preferably, the time window constraint analysis layer includes a first channel and a second channel. The first channel calculates the time delay probability of the input path data, and the second channel generates a time window distribution heat map after segmented clustering of the path. The heat map is weightedly superimposed with the delay probability to output the time constraint correction coefficient.

[0008] Preferably, during the second-level optimization process of the mixed integer programming, it is necessary to calculate the resource occupancy rate of the initial scheduling path generated in the first level, and adjust the order of the path nodes through a linear programming model.

[0009] Preferably, the cost constraint analysis layer performs segmented cost accounting on the input path data, including transportation cost, storage cost and loss cost, and generates a total cost constraint index through segmented weighted summation.

[0010] Preferably, the optimal parameter combination in the historical scheduling data is loaded when initializing the model parameters in step S3.

[0011] Preferably, the deviation loss calculation in step S5 adopts a weighted mean square error function, the matching loss calculation adopts a cosine similarity function, and the back propagation process uses an adaptive momentum optimization algorithm to update parameters.

[0012] Preferably, the multi-objective optimization scheduling model includes a real-time feedback correction module, which dynamically updates the demand forecast map and inventory status map by monitoring the actual transportation rate and inventory consumption rate of the logistics nodes, and inputs the updated data into the path correction module to regenerate the scheduling path planning map.

[0013] Preferably, the multi-stage optimization module is connected to a multi-dimensional evaluation layer, which performs a three-dimensional scoring of time efficiency, resource utilization and cost-effectiveness on the scheduling path planning diagram, and selects the scheduling plan with the highest comprehensive score through a weighted decision matrix.

[0014] Compared with the prior art, the present invention has the following beneficial effects: In terms of data processing and utilization, by acquiring real-time order and inventory data for medical consumables and performing standardized preprocessing, the problems of data chaos and inconsistent formats are effectively resolved. The demand forecast and inventory status diagrams generated on this basis can clearly and intuitively display demand trends and inventory dynamics, providing accurate data support for subsequent scheduling decisions. Compared with traditional simple data recording methods, this in-depth data processing and visualization greatly enhances the data's utility value and makes scheduling decisions more scientific and reasonable. The construction of a multi-objective optimization scheduling model is one of the core innovations of this method. This model takes mixed integer programming as its core and combines linear programming with dynamic programming to construct a two-layer decision network, breaking through the limitations of traditional single-objective optimization. The first layer generates the initial scheduling path through dynamic programming, providing a basic framework for subsequent optimization; the second layer performs multi-stage optimization of the initial path through linear programming, using a two-level path correction module to analyze the two key dimensions of time window constraints and cost constraints, and weightedly integrates the results to achieve refined adjustments to the scheduling path. At the same time, path smoothing, resource allocation normalization, and threshold activation operations performed after the multi-stage optimization module further enhance the feasibility and rationality of the scheduling path. This multi-objective, multi-level optimization mechanism can comprehensively consider multiple key factors such as transportation time, cost, and resource utilization. The resulting scheduling plan is more in line with actual needs and effectively balances the interests of all parties. By loading the optimal parameter combination from historical scheduling data during initialization, a solid foundation is laid for the model's efficient operation. During operation, the model uses a weighted mean square error function to calculate bias loss and a cosine similarity function to calculate matching loss. An adaptive momentum optimization algorithm is used for backpropagation parameter updates, enabling the model to continuously learn and optimize based on actual conditions, gradually improving the accuracy and effectiveness of scheduling solutions. This parameter optimization mechanism endows the model with powerful adaptability, enabling it to maintain strong performance in complex and changing supply chain environments. The setting of the real-time feedback correction module is another highlight of this method. By monitoring the actual transportation rate and inventory consumption rate of logistics nodes in real time, this module can promptly capture dynamic changes in the supply chain and dynamically update the demand forecast map and inventory status map accordingly. The updated data is input into the path correction module again to regenerate the scheduling path planning map, realizing the dynamic adjustment of the scheduling plan. This real-time feedback mechanism enables the scheduling plan to quickly adapt to environmental changes, effectively deal with uncertainties in logistics distribution, ensure that medical consumables can be delivered to the required location in a timely and accurate manner, and improve the response speed and reliability of the supply chain. The introduction of a multi-dimensional evaluation layer provides a scientific evaluation system for scheduling solution selection. This layer scores scheduling path plans based on three key dimensions: time efficiency, resource utilization, and cost-effectiveness. Using a weighted decision matrix, the scheduling solution with the highest overall score is selected. This comprehensive and objective evaluation method avoids the limitations of single-metric evaluation, enabling the selection of the optimal scheduling solution and maximizing the overall efficiency of the supply chain. Furthermore, through intelligent scheduling and dynamic inventory adjustments within the medical consumables supply chain, this method can effectively reduce inventory backlogs, lowering capital tie-up and the risk of loss. It also avoids delays in medical treatment caused by insufficient inventory, ensuring the continuity and stability of medical services. In the long term, this will help improve the management and operational efficiency of the entire medical consumables supply chain, promoting the healthy and sustainable development of the medical industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a working principle diagram of the medical consumables supply chain intelligent scheduling method according to the present invention; Figure 2 This is a working principle diagram of the multi-objective optimization scheduling model; Figure 3 This is a diagram showing the working principle of the time window constraint analysis layer. DETAILED DESCRIPTION

[0016] 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.

[0017] See also Figures 1 to 3 The present invention provides a method for intelligent scheduling of a medical consumables supply chain, and the specific implementation steps are as follows: First, execute step S1 to connect with relevant business platforms through the medical consumables supply chain system to obtain real-time order data and inventory data. These data cover the order time, required quantity, delivery address, type of consumables, as well as the existing inventory quantity and storage location of each warehouse. After obtaining the data, standardize and preprocess the data according to a unified data format standard, such as standardizing the date format and numerical unit in the data. Based on the preprocessed data, use the data analysis algorithm to generate a demand forecast map and an inventory status map. The demand forecast map shows the demand trend of various medical consumables in different time periods, and the inventory status map shows the real-time inventory level of different medical consumables in each warehouse.

[0018] Next, the multi-objective optimization scheduling model in step S2 is constructed. This model comprises a two-layer decision network based on linear programming and dynamic programming, with mixed integer programming at its core. The first layer of the model utilizes a dynamic programming algorithm to generate an initial scheduling path, taking into account various order attributes and constraints. The second layer uses linear programming to perform multi-stage optimization of this initial path. This multi-stage optimization module consists of two serially connected path correction modules. Each path correction module feeds the input path data into the time window constraint analysis layer and the cost constraint analysis layer, and then outputs a weighted fusion of the analysis results.

[0019] In step S3, the parameters of the multi-objective optimization scheduling model are initialized, including order priority weights and inventory turnover coefficients. During the initialization process, the optimal parameter combination is loaded from historical scheduling data. This historical scheduling data contains relevant data on the scheduling of medical consumables under different circumstances. The optimal parameter combination obtained by analyzing and summarizing this data can provide more reasonable parameter settings for the initial operation of the model.

[0020] Then, step S4 is executed, inputting the order data, demand forecast map, and inventory status map preprocessed in step S1 into the multi-objective optimization scheduling model. The model uses forward calculations based on its algorithmic logic and pre-set parameters to generate a scheduling path planning map and an inventory allocation map. The scheduling path planning map defines the specific transportation routes for medical consumables from each warehouse to each demand point, while the inventory allocation map determines the quantity of medical consumables allocated by each warehouse to different demand points.

[0021] In step S5, the dispatch path plan generated in step S4 is compared with the actual logistics path. The deviation loss is calculated using a weighted mean square error function, which considers the impact of different factors on path deviation and assigns corresponding weights. Simultaneously, the inventory allocation map is compared with the inventory status map, and the matching loss is calculated using a cosine similarity function to measure the degree of match between the inventory allocation plan and the actual inventory status. Based on the calculated deviation and matching losses, the model parameters are updated through backpropagation using an adaptive momentum optimization algorithm.

[0022] Then, step S6 is executed, and steps S4 and S5 are repeated to continuously adjust the model parameters until a preset number of iterations is reached.

[0023] Finally, step S7 is executed to output the scheduling plan and inventory dynamic adjustment strategy for medical consumables based on the optimized multi-objective optimization scheduling model. The scheduling plan specifies the specific transportation route and the shipment quantity of each warehouse, and the inventory dynamic adjustment strategy guides how to adjust the inventory level according to demand and inventory changes. Example 1

[0024] After the multi-objective optimization scheduling model completes the processing of the multi-stage optimization module, it enters the key link of generating the final scheduling path planning map. In the path smoothing processing stage, a smoothing algorithm based on curve fitting is used to operate on the path data. In the actual transportation scheduling of medical consumables, the initially generated scheduling path may have many sharp turns due to the discreteness of the algorithm calculation. These turns will not only increase the time cost and operation complexity during the transportation process, but may also affect the transportation efficiency. The curve fitting algorithm constructs a suitable curve function for the node coordinate data in the path, and converts the original broken line path into a smooth curve by interpolating and approximating the path nodes. For example, for a transportation route containing multiple warehouses and demand points, the algorithm will use the spline curve fitting method based on the latitude and longitude coordinates of each node to make the path transition between adjacent nodes more natural and smooth, reducing unnecessary route detours and turns.

[0025] During resource allocation normalization, data on various resources involved in a route, such as vehicle loading space, transportation time, and labor input, are normalized. Because data for different resource types vary in scale and units, effective analysis and comparison cannot be performed directly using the raw data. Normalization first determines the range of values for each resource. Then, using a specific mathematical transformation formula, all resource data is mapped to a value between 0 and 1. For example, vehicle loading space and transportation time are two resources. Loading space may be measured in cubic meters, with values ranging from tens to hundreds. Transportation time, on the other hand, may be measured in hours, with values ranging from a few to dozens of hours. By normalizing these two data types, which have different scales and units, they are brought to the same standard. This allows the model to more fairly comprehensively evaluate and optimize the allocation of different resources, facilitating subsequent decision-making and analysis.

[0026] In the threshold activation phase, reasonable threshold conditions are set for the data after path smoothing and resource allocation normalization based on the actual business needs and experience of the medical consumables supply chain. For path-related data, set thresholds such as the total length of the path, the maximum time threshold of a single transport segment in the path, etc.; for resource allocation data, set resource utilization thresholds, unit transport volume resource consumption thresholds, etc. When the data meets the preset threshold conditions, this part of the data is activated and included in the generation process of the final scheduling path planning map; if the threshold conditions are not met, the data is adjusted or recalculated. For example, if the total length of a scheduling path exceeds the set longest path threshold, the system will re-examine the node connection method of the path and try to adjust the transportation route until the threshold requirements are met. Finally, a final scheduling path planning map that meets the actual needs of medical consumables transportation and is scientific and reasonable is generated, providing accurate and effective guidance for the actual medical consumables transportation scheduling. Example 2

[0027] During the operation of the multi-objective optimization scheduling model, the time window constraint analysis layer uses a dual-channel collaborative working mechanism to refine the scheduling path. The first channel calculates the time delay probability using a Bayesian network-based prediction model, whose parameter settings fully consider the special properties of medical consumables transportation. The model classifies weather conditions in historical transportation data into different levels: sunny, cloudy, rainy, and snowy. The traffic congestion index is divided into four levels based on road capacity: smooth, lightly congested, moderately congested, and severely congested. Different weight coefficients are assigned to different types of medical consumables, such as emergency, conventional, and refrigerated. During the actual calculation process, after inputting the specific transportation time, departure point, and destination information of a particular route, the model first calculates the frequency of delays under different weather and traffic conditions based on historical data. Then, combined with the transportation time of the current route, it calculates the probability distribution of the weather and traffic conditions that may be encountered at that time. Finally, based on the characteristic weights of different medical consumables, the time delay probability of the route is comprehensively calculated.

[0028] The path segment clustering of the second channel uses a density peak clustering algorithm, which can adaptively identify key nodes and natural segments in the path. When analyzing the transportation path of medical consumables, the algorithm calculates the similarity and density distribution between nodes based on multi-dimensional data such as the geographic coordinates, transportation time, and transportation volume of each node on the path. For a transportation path that includes multiple warehouses and medical institutions, the algorithm first identifies key nodes with large transportation volumes and long transportation times. These nodes are usually large warehouses or key medical institutions. Then, based on the distance and transportation correlation between the nodes, the path is divided into several natural segments, each segment representing a relatively independent transportation unit. When generating a time window distribution heat map, a two-dimensional coordinate system is constructed with geographic coordinates as the horizontal axis and time as the vertical axis. The transportation demand intensity of each segment in different time periods is mapped to the coordinate system, and the intensity of the demand is represented by the depth of color, forming an intuitive time window distribution heat map.

[0029] When weighted superposition of the time window distribution heat map and the delay probability is performed, the system will dynamically adjust the weight coefficient according to the urgency of the transportation task. For the transportation tasks of emergency medical consumables, the system will automatically increase the weight of the delay probability, because the timely delivery of such consumables is crucial to the patient's life safety, and any delay may have serious consequences. For the transportation tasks of conventional medical consumables, the system will relatively reduce the weight of the delay probability and appropriately increase the weight of the time window distribution heat map to balance transportation efficiency and cost. Through this weighted superposition method, the system can comprehensively consider the time delay risk and time window constraints, and generate a correction coefficient that accurately reflects the path time constraint status. This correction coefficient is used to adjust and optimize the scheduling path to ensure that medical consumables can be accurately delivered to the destination within the specified time window. Example 3

[0030] During the second-level optimization process of the mixed integer programming (MIP) multi-objective optimization scheduling model, the resource utilization calculation phase conducts a detailed analysis of transportation vehicles and storage facilities for the initial scheduling route generated in the first level. This calculation is based on the rated loading capacity and maximum load capacity of different types of transportation vehicles (such as vans and cold chain transport vehicles), combined with the quantity, specifications, and packaging dimensions of the medical consumables planned to be loaded on the initial scheduling route. The space utilization ratio of each transportation vehicle along each transportation route segment is calculated. Furthermore, the operating time of the transportation vehicle is taken into account, and the time utilization ratio is calculated by comparing the transportation time with the total operating time of the transportation vehicle. For example, if a cold chain transport vehicle with a rated loading capacity of 30 cubic meters is required to load a total volume of refrigerated medical consumables of 20 cubic meters on a certain transportation route segment, its space utilization ratio is 20÷30. If the vehicle is available for 12 hours per day and the transportation segment takes 3 hours, the time utilization ratio is 3÷12.

[0031] Regarding the resource utilization rate of storage facilities, the warehouse space utilization rate is calculated based on information such as the total storage area, number of shelf layers, and effective storage area per layer of each warehouse, combined with the planned storage quantity and storage method of medical consumables in the initial scheduling path. For example, if a warehouse has a total storage area of 1,000 square meters and an effective shelf storage area of 800 square meters, and the planned storage of medical consumables requires 400 square meters of shelf space, the warehouse space utilization rate is 400 ÷ 800. Furthermore, the inventory turnover cycle is taken into account, and the estimated storage time of medical consumables is compared with the warehouse's ideal inventory turnover cycle to derive the resource utilization rate for inventory turnover.

[0032] After completing the resource utilization calculation, the linear programming model uses resource utilization as the core constraint and adjusts the order of path nodes in combination with factors such as order priority and demand urgency. The linear programming model regards each transportation node as a decision variable and constructs a mathematical model that includes constraints such as resource utilization restrictions and order delivery time restrictions. During the solution process, the transportation flow and storage arrangements of medical consumables are changed by adjusting the connection order between nodes. For example, when the space resource utilization rate of a transportation tool on a certain section of the route is too high, the model will try to adjust the transportation tasks of some goods to other transportation tools, or change the transportation route to give priority to meeting the transportation needs of high-priority orders, while balancing the resource utilization of each transportation tool and storage facility, making the overall resource utilization more reasonable, reducing the increase in transportation costs and efficiency loss caused by excessive concentration or uneven distribution of resources, realizing the optimization and reorganization of medical consumables transportation routes, and improving the overall operational efficiency of the supply chain. Example 4

[0033] The cost constraint analysis layer performs segmented cost accounting on the input route data, specifically dividing the transport route into different segments based on factors such as geographic region and transport mode conversion points. Within each segment, transportation costs, storage costs, and loss costs are calculated separately.

[0034] When calculating transportation costs, factors such as transportation distance, type of transportation vehicle, and unit price are considered. For road transportation, the unit price is determined based on the type of vehicle (such as small truck, large truck), fuel price, tolls, etc.; for air transportation, the unit price is determined based on flight fees, cargo weight and volume, etc. (Unit: km), the unit price of transportation is (Unit: Yuan / km), then the transportation cost of this segment is ,in represents the transportation cost, is the actual transport distance of the segment, It is the unit transportation price determined based on the mode of transportation and other relevant factors.

[0035] In terms of warehousing cost accounting, warehouse rental fees, inventory management fees, etc. are taken into consideration. If a warehouse charges rental fees on a daily basis, the daily rental price is (Unit: Yuan / day), the number of days medical consumables are stored in this warehouse is (Unit: day), and the daily inventory management cost (including personnel management, equipment maintenance, etc.) is (Unit: Yuan / day), then the storage cost of the warehouse for storing this batch of medical consumables is .

[0036] The loss cost calculation is aimed at the damage and expiration of medical consumables during transportation and storage. For fragile medical consumables, the damage quantity is estimated based on the historical damage rate and the current transportation quantity; for medical consumables with shelf life requirements, the expired quantity is calculated based on the storage time and shelf life. Assume that the initial quantity of a certain type of medical consumables is , the unit value is (Unit: Yuan), the estimated loss is , then the loss cost of this type of medical consumables is .

[0037] After calculating the transportation cost, storage cost and loss cost of each segment, the total cost constraint index is generated by weighted summation of the segments. segments, and the cost of each segment is , and the corresponding weights are , and satisfies The weight is determined based on a comprehensive assessment of factors such as the transportation distance, transportation difficulty, and cargo value of each segment. Total cost constraint indicator ,in The total cost constraint indicator representing the entire transportation path is used to measure the cost of different scheduling paths, providing a quantitative basis for the model to select the scheduling path with the optimal cost, thereby achieving effective cost control in the scheduling of medical consumables supply chains. Example 5

[0038] The real-time feedback correction module in the multi-objective optimization scheduling model establishes a stable data connection with the GPS positioning system and warehouse management system of the logistics node to continuously monitor the actual transportation rate and inventory consumption rate. In the logistics transportation link, the system obtains the location information, mileage, driving time and other data of the transport vehicle in real time, and calculates the mileage traveled per unit time to obtain the actual transportation rate. For example, the mileage traveled by the transport vehicle in a certain period of time is kilometers, and the driving time is hours, the actual transport rate during this period km / h.

[0039] In terms of inventory management, the number and time of medical consumables leaving the warehouse are counted in real time to calculate the inventory consumption rate. When medical consumables leave the warehouse, the number of medical consumables leaving the warehouse is recorded. and the corresponding time interval , inventory consumption rate Once the actual transportation rate or inventory consumption rate is detected to deviate from the model's initial prediction, the real-time feedback correction module will initiate the data update process.

[0040] Specifically, this module re-analyzes and adjusts the demand forecast chart and inventory status chart based on actual monitoring data, combined with historical demand data and market trends. When updating the demand forecast chart, the module uses time series analysis methods to incorporate new actual demand data into the analysis scope, rebuilds the demand forecast model, and predicts the demand quantity of different types of medical consumables in different time periods in the future, adjusting the trend and value of the demand forecast curve. For the inventory status chart, the real-time inventory levels of different medical consumables in each warehouse are updated based on the latest inventory consumption rate and actual inventory quantity, and the inventory values and warning signs on the inventory status chart are re-marked.

[0041] The updated data will be input into the path correction module in a timely manner. After receiving the new data, the path correction module will re-analyze the scheduling path based on the algorithm logic of the multi-objective optimization scheduling model. The module first checks whether the existing scheduling path can still meet the new demand and inventory situation. If not, the node sequence, transportation mode selection, warehouse allocation strategy, etc. in the path will be recalculated and adjusted according to the algorithms of the time window constraint analysis layer and the cost constraint analysis layer. For example, if the inventory consumption rate of medical consumables in a certain area accelerates, resulting in lower-than-expected inventory levels, the path correction module may adjust the transportation route, give priority to allocating goods from nearby warehouses, or change the transportation mode to improve transportation efficiency, thereby regenerating a scheduling path planning map that conforms to the current actual situation, ensuring that the scheduling plan of the medical consumables supply chain can adapt to actual business changes in real time and maintain the efficient and stable operation of the supply chain.

[0042] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0043] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent scheduling of medical consumables supply chain, characterized in that: The following steps are involved: Step S1: Acquire real-time order data and inventory data of medical consumables, perform standardization preprocessing on the data, and generate a demand forecast graph and an inventory status graph; Step S2: constructing a multi-objective optimization scheduling model, wherein the model includes a two-layer decision network based on linear programming and dynamic programming; Step S3: Initialize the parameters of the multi-objective optimization scheduling model, including order priority weights and inventory turnover coefficients; Step S4: input the order data, demand forecast graph, and inventory status graph pre-processed in step S1 into the multi-objective optimization scheduling model, and generate a scheduling path planning graph and an inventory allocation graph through forward calculation; Step S5: Calculate the deviation loss between the scheduling path planning map generated in step S4 and the actual logistics path, calculate the matching loss between the inventory allocation map and the inventory status map, and backpropagate to update the model parameters; Step S6, repeating steps S4 and S5 to a preset number of iterations; Step S7: Based on the optimized multi-objective optimization scheduling model, output the scheduling plan and inventory dynamic adjustment strategy for medical consumables.

2. The intelligent scheduling method for the medical consumables supply chain according to claim 1, characterized in that: The multi-objective optimization scheduling model is based on mixed integer programming. The first layer of the multi-objective optimization scheduling model generates an initial scheduling path through dynamic programming, and the second layer performs multi-stage optimization on the initial path through linear programming. The multi-stage optimization module is composed of two levels of path correction modules connected in series. The path correction module inputs the input path data into the time window constraint analysis layer and the cost constraint analysis layer respectively, and outputs the weighted fusion of the analysis results.

3. The intelligent scheduling method for the medical consumables supply chain according to claim 2, characterized in that: The multi-objective optimization scheduling model performs path smoothing, resource allocation normalization and threshold activation in sequence after the multi-stage optimization module to generate a final scheduling path planning diagram.

4. The intelligent scheduling method for the medical consumables supply chain according to claim 2, characterized in that: The time window constraint analysis layer includes a first channel and a second channel. The first channel calculates the time delay probability of the input path data, and the second channel generates a time window distribution heat map after segmented clustering of the path. The heat map is weightedly superimposed with the delay probability to output the time constraint correction coefficient.

5. The intelligent scheduling method for the medical consumables supply chain according to claim 2, characterized in that: During the second-level optimization process of the mixed integer programming, the resource occupancy rate of the initial scheduling path generated in the first level needs to be calculated, and the order of the path nodes needs to be adjusted through a linear programming model.

6. The intelligent scheduling method for the medical consumables supply chain according to claim 2, characterized in that: The cost constraint analysis layer performs segmented cost accounting on the input path data, including transportation cost, storage cost and loss cost, and generates a total cost constraint index through segmented weighted summation.

7. The intelligent scheduling method for the medical consumables supply chain according to claim 1, characterized in that: When initializing the model parameters in step S3, the optimal parameter combination in the historical scheduling data is loaded.

8. The intelligent scheduling method for the medical consumables supply chain according to claim 1, characterized in that: The deviation loss calculation in step S5 adopts the weighted mean square error function, the matching loss calculation adopts the cosine similarity function, and the back propagation process uses the adaptive momentum optimization algorithm to update the parameters.

9. The intelligent scheduling method for the medical consumables supply chain according to claim 1, characterized in that: The multi-objective optimization scheduling model includes a real-time feedback correction module, which dynamically updates the demand forecast map and inventory status map by monitoring the actual transportation rate and inventory consumption rate of the logistics nodes, and inputs the updated data into the path correction module to regenerate the scheduling path planning map.

10. The intelligent scheduling method for the medical consumables supply chain according to claim 1, characterized in that: The multi-stage optimization module is connected to a multi-dimensional evaluation layer, which performs a three-dimensional scoring of time efficiency, resource utilization and cost-effectiveness on the scheduling path planning map, and selects the scheduling plan with the highest comprehensive score through a weighted decision matrix.

Images