Neural network-based medicine cold chain inventory optimization method and system

Abstract

The invention discloses a medicine cold chain inventory optimization method and system based on a neural network, and the method comprises the steps: S1, collecting related data of medicine cold chain inventory, including inventory, inventory period, sales volume and demand prediction; s2, modeling by using the data, and monitoring the current situation of the inventory and optimizing the space through a model; s3, inventory checking is carried out, whether the actual inventory is matched with the recorded inventory or not is checked, and the actual quality and the validity period of the inventory are evaluated; and S4, optimizing demand prediction, predicting the demand quantity and the sales trend of the medicine products, and reducing the inventory fluctuation, improving the inventory turnover rate and optimizing the inventory management by optimizing the demand prediction. By training the neural network model, modeling and prediction of a complex system can be realized, the neural network is applied to medicine cold chain inventory optimization, historical data and real-time data can be fully utilized, the accuracy of demand prediction is improved, an inventory management strategy is optimized, the inventory cost is reduced, and the competitiveness of an enterprise is improved.

Description

Technical Field

[0001] The present invention belongs to the field of inventory optimization, and more specifically, it particularly relates to a pharmaceutical cold chain inventory optimization method based on a neural network. At the same time, the present invention also relates to a pharmaceutical cold chain inventory optimization system based on a neural network. Background Art

[0002] With the rapid development of the pharmaceutical industry, pharmaceutical cold chain logistics plays a vital role in the circulation of drugs. Pharmaceutical cold chain logistics involves multiple links such as production, storage, transportation, and sales of drugs, which has a significant impact on the quality, safety, and effectiveness of drugs. Therefore, optimizing pharmaceutical cold chain inventory management, improving inventory turnover, and reducing inventory costs are of great significance to pharmaceutical companies.

[0003] Traditional pharmaceutical cold chain inventory management methods mainly rely on manual experience and simple mathematical models for prediction and decision-making. This method often has large errors and uncertainties. Specifically, the existing technology has the following shortcomings:

[0004] Traditional methods often rely on managers’ experience to make decisions, which is easily affected by personal subjective factors, resulting in inaccurate predictions and decisions;

[0005] Limited data processing capabilities: Traditional methods are unable to process large-scale, multi-dimensional data and are unable to fully utilize the information contained in historical and real-time data;

[0006] Lack of dynamic adjustment mechanism: Traditional inventory management methods usually adopt static inventory strategies, which are difficult to dynamically adjust according to changes in market demand and supply chain environment;

[0007] On the basis of the above, we propose a neural network-based pharmaceutical cold chain inventory optimization method and system to effectively solve the shortcomings of the above-mentioned existing technologies. Summary of the invention

[0008] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose a pharmaceutical cold chain inventory optimization method and system based on neural network. By training the neural network model, it is possible to achieve modeling and prediction of complex systems. The application of neural network to pharmaceutical cold chain inventory optimization can make full use of historical data and real-time data, improve the accuracy of demand forecasting, thereby optimizing inventory management strategies, reducing inventory costs, and improving the competitiveness of enterprises.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] The pharmaceutical cold chain inventory optimization method based on neural network includes the following steps:

[0011] S1. Collect relevant data on pharmaceutical cold chain inventory, including inventory volume, inventory cycle, sales volume, and demand forecast;

[0012] S2. Use data to build models, monitor the current status of inventory and optimize space through models;

[0013] S3. Conduct inventory counts to check whether the actual inventory matches the recorded inventory, and evaluate the actual quality and expiration date of the inventory;

[0014] S4. Optimize demand forecasts, predict the demand and sales trends of pharmaceutical products, reduce inventory fluctuations, improve inventory turnover, and optimize inventory management by optimizing demand forecasts.

[0015] Preferably, step S1 specifically comprises:

[0016] S11. Identify the types of data that need to be collected, develop a detailed data collection plan, and collect data from the past 12 months based on the data types;

[0017] S12. Extract the required data from various data sources, ensure the integrity and accuracy of the data, perform preliminary cleaning and verification of the data, and eliminate erroneous and inconsistent data;

[0018] S13. Store the collected data in a data warehouse and establish a data management mechanism to ensure data accessibility, security and privacy protection;

[0019] S14. Further clean and process the collected data, including removing duplicate data, filling missing values, and standardizing data formats;

[0020] S15. Integrate data from different sources to form a unified data set and back up the data set.

[0021] Preferably, step S2 specifically comprises:

[0022] S21. According to business requirements and data characteristics, select appropriate features as the basis for clustering, and select clustering algorithm modeling according to data characteristics and clustering goals;

[0023] S22. Use the selected clustering algorithm to perform cluster analysis on the data, classify similar data points into the same category, and analyze the clustering results to identify the characteristics and differences of different categories;

[0024] S23. Develop targeted inventory management strategies based on the clustering results. For products with high inventory levels but low turnover rates, reduce purchases or adjust sales strategies. For products with large sales fluctuations, strengthen demand forecasting and inventory control.

[0025] Preferably, the clustering algorithm modeling is selected according to the characteristics of the data and the clustering target, specifically, the data set is divided into k clusters, each cluster is represented by a centroid, and the square error of the sample points in the cluster is minimized by iteratively updating the position of the centroid. The process is:

[0026] S211, initialization, randomly select k initial centroids μ1,μ2,…,μ k ;

[0027] S212, assign sample points, assign each sample point xi to the nearest centroid μ j In the cluster represented by; when allocating sample points, Euclidean distance is usually used to measure the similarity between sample points and centroids, and the expression is:

[0028]

[0029] In the formula, d(x i ,μ j ) represents the sample point x i With the center of mass μ j The distance between i is the sample point, μ j is the centroid, n is the number of features;

[0030] S213, update the centroid, recalculate the centroid of each cluster, the expression is:

[0031] |C j | represents cluster C j The number of sample points in ;

[0032] S214, repeat steps S212 and S213 until the centroid no longer changes or a preset number of iterations is reached.

[0033] Preferably, step S3 specifically comprises:

[0034] S31. Determine the time range for inventory taking and notify warehouse management personnel in advance;

[0035] S32. Start the inventory at the scheduled time, ensure that all personnel are in place, scan the inventory items one by one, and record the actual inventory quantity;

[0036] Compare the actual inventory quantity with the inventory quantity recorded in the system. For any mismatch, make detailed records and find out the reasons.

[0037] S33. Check the appearance, packaging and labels of inventory goods to ensure that they are not damaged or expired. For perishable and deteriorating goods, pay special attention to their expiration dates and storage conditions.

[0038] Preferably, in step S4, the model is used to optimize the demand forecast, and the model expression is:

[0039] y=β0+β1x1+β2x2+…+β p x p +∈;

[0040] In the formula, y is the dependent variable, i.e., demand and sales trend, β0 is the intercept term, β1, β2, .., β p are the regression coefficients, x1, x2, ..., x p is the independent variable and E is the error term.

[0041] Preferably, the optimization of inventory management in step S4 also needs to be performed in combination with a weighted algorithm, specifically:

[0042] 1) Integrate the shelf life, distance to the destination and product type information of related products;

[0043] 2) Convert each factor into a numerical value. The shelf life is calculated based on the remaining days, the distance to the destination is directly calculated using kilometers, and the product type is assigned different scores based on market demand and value.

[0044] 3) Assign appropriate weight coefficients to each factor based on business needs and strategies;

[0045] 4) Calculate the weighted score of each product through weighted calculation, sort all products according to the calculated weighted score, and give priority to products with high scores to optimize inventory turnover and reduce waste.

[0046] Preferably, after weighted calculation, if the same type of products have the same score, they will be shipped out using the first-in-first-out principle, and different types of products will be multiplied by a constant term on the basis of the weighted score as the final weighted score, and the constant term is the score assigned to the product type based on market demand and value.

[0047] Preferably, the weighted score calculation formula for each product calculated by weighted calculation is:

[0048] Weighted Score = w SL ×SL+w DD ×DD+w PT ×PT;

[0049] In the formula, SL is the shelf life of the product, which is obtained by quantifying the remaining days; DD is the distance to the destination, which is obtained by quantifying the distance in kilometers; PT is the weight of the product type, with a score range of 1-5;

[0050] The quantitative expression of the product's shelf life is:

[0051]

[0052] The quantitative expression of the distance in kilometers to the destination is:

[0053] DD = actual kilometres received / quantified benchmark kilometres;

[0054] Among them, the calculation formula for the quantitative benchmark kilometers is:

[0055] Quantitative benchmark mileage = (actual mileage received × 2) × 0.618; that is, the total round-trip mileage of the receiving location multiplied by the golden ratio;

[0056] The weights of product types are:

[0057] Conventional medicines, used to treat common diseases, usually not required for emergency use, and with a large number of alternatives, are scored 1;

[0058] Commonly used drugs, which are relatively common but may need to be given priority in some cases, including antibiotics and conventional analgesics, with a score of 2;

[0059] Important drugs are very important for certain specific diseases or patient groups and may require special attention, including antihypertensive drugs and hypoglycemic drugs, with a score of 3;

[0060] Critical drugs, which play a key role in emergency situations or in the treatment of certain serious diseases, including emergency hemostatic drugs and cardiovascular emergency drugs, are scored 4;

[0061] Emergency medicines, which are essential in medical emergencies and have the highest priority, include adrenaline, a powerful, addictive painkiller, and are rated 5.

[0062] A pharmaceutical cold chain inventory optimization system based on a neural network, the system is used to implement the above method, including:

[0063] The data collection and preprocessing module is used to collect and preprocess the collected inventory data;

[0064] The model building and analysis module is used to build models based on the data preprocessed by the data collection and preprocessing module, and formulate targeted inventory management strategies based on the model clustering results;

[0065] The inventory counting and quality assessment module is used to assess the actual quality and validity period of the inventory based on the data preprocessed by the data collection and preprocessing module;

[0066] The demand forecasting and inventory optimization module is used to optimize demand forecasting and reduce inventory fluctuations by optimizing demand forecasting, improve inventory turnover, and optimize inventory management.

[0067] Technical effects and advantages of the present invention: Compared with the prior art, the pharmaceutical cold chain inventory optimization method and system provided by the present invention have the following beneficial effects:

[0068] By leveraging the powerful learning and forecasting capabilities of neural networks, it is possible to more accurately forecast the demand and sales trends of pharmaceutical products, which helps reduce inventory backlogs or out-of-stock situations caused by inaccurate forecasts, thereby reducing inventory costs and improving inventory turnover. By modeling and analyzing inventory data, it is possible to identify the characteristics and differences of different categories of products and develop targeted inventory management strategies.

[0069] By regularly taking inventory, checking whether the actual inventory matches the recorded inventory, and evaluating the actual quality and expiration date of the inventory, the accuracy and reliability of the inventory data can be ensured. At the same time, this also helps to promptly discover and resolve inventory anomalies and avoid potential losses. By integrating the shelf life, distance to the destination, and product type information of related products, and combining them with weighted algorithms for calculation, collaborative optimization between various links in the supply chain can be achieved. At the same time, by real-time monitoring of inventory status and early warning, potential risk points can be discovered in a timely manner and appropriate measures can be taken to deal with them. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 This is a flow chart of the pharmaceutical cold chain inventory optimization method based on neural network of the present invention;

[0071] Figure 2 This is a flow chart of relevant data collected from pharmaceutical cold chain inventory in an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0073] The present invention provides a pharmaceutical cold chain inventory optimization method and system based on a neural network, which realizes the optimization management of pharmaceutical cold chain inventory through the steps of collecting, modeling, analyzing, counting and demand forecasting relevant data of pharmaceutical cold chain inventory. The method combines the powerful learning ability of neural networks with the advantages of traditional inventory management methods, can effectively improve the efficiency and accuracy of pharmaceutical cold chain inventory management, and bring significant economic benefits to pharmaceutical companies.

[0074] like Figure 1 As shown in the figure, the pharmaceutical cold chain inventory optimization method based on neural network is as follows:

[0075] S1. Collect relevant data on pharmaceutical cold chain inventory, including inventory quantity, inventory cycle, sales volume, and demand forecast; Step S1 is specifically as follows:

[0076] S11. Identify the types of data that need to be collected, develop a detailed data collection plan, and collect data from the past 12 months based on the data types;

[0077] S12. Extract the required data from various data sources, ensure the integrity and accuracy of the data, perform preliminary cleaning and verification of the data, and eliminate erroneous and inconsistent data;

[0078] S13. Store the collected data in a data warehouse and establish a data management mechanism to ensure data accessibility, security and privacy protection;

[0079] S14. Further clean and process the collected data, including removing duplicate data, filling missing values, and standardizing data formats;

[0080] S15. Integrate data from different sources to form a unified data set and back up the data set.

[0081] S2. Use data to build a model, and use the model to monitor the current inventory status and optimization space; Step S2 is specifically as follows:

[0082] S21. According to business requirements and data characteristics, select appropriate features as the basis for clustering, and select clustering algorithm modeling according to data characteristics and clustering goals;

[0083] The clustering algorithm modeling is selected according to the characteristics of the data and the clustering target. Specifically, the data set is divided into k clusters, each cluster is represented by a centroid, and the square error of the sample points in the cluster is minimized by iteratively updating the position of the centroid. The process is:

[0084] S211, initialization, randomly select k initial centroids μ1,μ2,…,μ k ;

[0085] S212, assign sample points, assign each sample point xi to the nearest centroid μ j In the cluster represented by; when allocating sample points, Euclidean distance is usually used to measure the similarity between sample points and centroids, and the expression is:

[0086]

[0087] In the formula, d(x i ,μ j ) represents the sample point x i With the center of mass μ j The distance betweeni is the sample point, μ j is the centroid, n is the number of features;

[0088] S213, update the centroid, recalculate the centroid of each cluster, the expression is:

[0089] |C j | represents cluster C j The number of sample points in ;

[0090] S214, repeating steps S212 and S213 until the centroid no longer changes or a preset number of iterations is reached;

[0091] S22. Use the selected clustering algorithm to perform cluster analysis on the data, classify similar data points into the same category, and analyze the clustering results to identify the characteristics and differences of different categories;

[0092] S23. Develop targeted inventory management strategies based on the clustering results. For products with high inventory levels but low turnover rates, reduce purchases or adjust sales strategies. For products with large sales fluctuations, strengthen demand forecasting and inventory control.

[0093] S3: Conduct an inventory count to check whether the actual inventory matches the recorded inventory, and evaluate the actual quality and validity period of the inventory; Step S3 is specifically as follows:

[0094] S31. Determine the time range for inventory taking and notify warehouse management personnel in advance;

[0095] S32. Start the inventory at the scheduled time, ensure that all personnel are in place, scan the inventory items one by one, and record the actual inventory quantity;

[0096] Compare the actual inventory quantity with the inventory quantity recorded in the system. For any mismatch, make detailed records and find out the reasons.

[0097] S33. Check the appearance, packaging and labels of inventory goods to ensure that they are not damaged or expired. For perishable and deteriorating goods, pay special attention to their expiration dates and storage conditions.

[0098] S4. Optimize demand forecasting, predict the demand and sales trends of pharmaceutical products, reduce inventory fluctuations, improve inventory turnover, and optimize inventory management by optimizing demand forecasting;

[0099] The optimization of inventory management in step S4 also needs to be performed in combination with a weighted algorithm, specifically:

[0100] 1) Integrate the shelf life, distance to the destination and product type information of related products;

[0101] 2) Convert each factor into a numerical value. The shelf life is calculated based on the remaining days, the distance to the destination is directly calculated using kilometers, and the product type is assigned different scores based on market demand and value.

[0102] 3) Assign appropriate weight coefficients to each factor based on business needs and strategies;

[0103] 4) Calculate the weighted score of each product through weighted calculation, sort all products according to the calculated weighted score, and give priority to products with high scores to optimize inventory turnover and reduce waste; after weighted calculation, if the same type of products have the same score, they will be shipped out on a first-in-first-out basis, and different types of products will be multiplied by a constant term on the basis of the weighted score to obtain the final weighted score, where the constant term is the score assigned to the product type according to market demand and value;

[0104] In step S4, the model is used to optimize the demand forecast, and the model expression is:

[0105] y=β0+β1x1+β2x2…+βpxxpr.

[0106] In the formula, y is the dependent variable, i.e., demand and sales trend, β0 is the intercept term, β1, β2, .., βp are regression coefficients, x1, x2, ..., xp are independent variables, and E is the error term;

[0107] Specifically, the weighted score calculation formula for each product is calculated by weighted calculation:

[0108] WeightedScore=wS L ×SLww D D× D )D+wPT×PT.

[0109] In the formula, SL is the shelf life of the product, which is obtained by quantifying the remaining days; DD is the distance to the destination, which is obtained by quantifying the distance in kilometers; PT is the weight of the product type, with a score range of 1-5;

[0110] The quantitative expression of the product's shelf life is:

[0111]

[0112] The quantitative expression of the distance in kilometers to the destination is:

[0113] DD = actual kilometres received / quantified benchmark kilometres;

[0114] Among them, the calculation formula for the quantitative benchmark kilometers is:

[0115] Quantitative benchmark mileage = (actual mileage received × 2) × 0.618; that is, the total round-trip mileage of the receiving location multiplied by the golden ratio;

[0116] The weights of product types are:

[0117] Conventional medicines, used to treat common diseases, usually not required for emergency use, and with a large number of alternatives, are scored 1;

[0118] Commonly used drugs, which are relatively common but may need to be given priority in some cases, including antibiotics and conventional analgesics, with a score of 2;

[0119] Important drugs are very important for certain specific diseases or patient groups and may require special attention, including antihypertensive drugs and hypoglycemic drugs, with a score of 3;

[0120] Critical drugs, which play a key role in emergency situations or in the treatment of certain serious diseases, including emergency hemostatic drugs and cardiovascular emergency drugs, are scored 4;

[0121] Emergency medicines, which are essential in medical emergencies and have the highest priority, include adrenaline, a powerful, addictive painkiller, and are rated 5.

[0122] For example, a pharmaceutical company needs to optimize its inventory management. They have five different types of drugs: regular drugs, commonly used drugs, important drugs, critical drugs, and emergency drugs. In order to manage inventory effectively, the company decided to use a weighted algorithm to evaluate the priority of each drug. The data is as follows:

[0123] Conventional drugs: SL = 0.8, DD = 0.8, PT = 1;

[0124] Commonly used drugs: SL = 0.6, DD = 0.75, PT = 2;

[0125] Important drugs: SL = 0.56, DD = 0.78, PT = 3;

[0126] Key drugs: SL = 0.77, DD = 0.85, PT = 4;

[0127] Emergency drugs: SL = 0.5, DD = 0.9, PT = 5;

[0128] Weighted score for each drug:

[0129] The weight coefficient of shelf life (SL) is 0.4;

[0130] The weight coefficient of the distance to the destination (DD) is 0.3;

[0131] The weight coefficient of product type (PT) is 0.3;

[0132] The weighted score is calculated as:

[0133] Weighted score = (SL × 0.4) + (DD × 0.3) + (PT × 0.3);

[0134] Conventional medicines

[0135] Weighted score = (0.8 × 0.4) + (0.8 × 0.3) + (1 × 0.3);

[0136] Weighted score = 0.32 + 0.24 + 0.3 = 0.86;

[0137] Commonly used drugs

[0138] Weighted score = (0.6 × 0.4) + (0.75 × 0.3) + (2 × 0.3);

[0139] Weighted score = 0.24 + 0.225 + 0.6 = 1.065;

[0140] Important medicines

[0141] Weighted score = (0.56 × 0.4) + (0.78 × 0.3) + (3 × 0.3);

[0142] Weighted score = 0.224 + 0.234 + 0.9 = 1.358;

[0143] Key drugs

[0144] Weighted score = (0.77 × 0.4) + (0.85 × 0.3) + (4 × 0.3);

[0145] Weighted score = 0.308 + 0.255 + 1.2 = 1.763;

[0146] Emergency medicines

[0147] Weighted score = (0.5 × 0.4) + (0.9 × 0.3) + (5 × 0.3);

[0148] Weighted score = 0.2 + 0.27 + 1.5 = 1.97;

[0149] Conclusion, emergency drugs have the highest weighted score, which indicates that they are of the highest importance in emergency medical situations and need to be given priority.

[0150] Next in line are critical medicines, which play a key role in treating serious diseases and therefore have a higher priority;

[0151] Important medicines, although very important for a specific disease or patient group, have a lower priority than emergency and critical medicines;

[0152] Commonly used drugs, while they may need priority in certain circumstances, are generally less urgent than the above categories and therefore are ranked fourth;

[0153] Conventional medicines are usually used to treat common diseases and have a large number of alternatives, so they have the lowest priority.

[0154] This embodiment also proposes a pharmaceutical cold chain inventory optimization system based on a neural network, and the system is used to implement the above method, such as Figure 2 As shown, including:

[0155] The data collection and preprocessing module is used to collect and preprocess the collected inventory data;

[0156] The model building and analysis module is used to build models based on the data preprocessed by the data collection and preprocessing module, and formulate targeted inventory management strategies based on the model clustering results;

[0157] The inventory counting and quality assessment module is used to assess the actual quality and validity period of the inventory based on the data preprocessed by the data collection and preprocessing module;

[0158] The demand forecasting and inventory optimization module is used to optimize demand forecasting and reduce inventory fluctuations by optimizing demand forecasting, improve inventory turnover, and optimize inventory management.

[0159] In summary, the present invention has the following beneficial effects:

[0160] By leveraging the powerful learning and forecasting capabilities of neural networks, it is possible to more accurately forecast the demand and sales trends of pharmaceutical products, which helps reduce inventory backlogs or out-of-stock situations caused by inaccurate forecasts, thereby reducing inventory costs and improving inventory turnover. By modeling and analyzing inventory data, it is possible to identify the characteristics and differences of different categories of products and develop targeted inventory management strategies.

[0161] By regularly taking inventory, checking whether the actual inventory matches the recorded inventory, and evaluating the actual quality and expiration date of the inventory, the accuracy and reliability of the inventory data can be ensured. At the same time, this also helps to promptly discover and resolve inventory anomalies and avoid potential losses. By integrating the shelf life, distance to the destination, and product type information of related products, and combining them with weighted algorithms for calculation, collaborative optimization between various links in the supply chain can be achieved. At the same time, by real-time monitoring of inventory status and early warning, potential risk points can be discovered in a timely manner and appropriate measures can be taken to deal with them.

[0162] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A pharmaceutical cold chain inventory optimization method based on neural network, characterized in that: The steps include: S1. Collect relevant data on pharmaceutical cold chain inventory, including inventory volume, inventory cycle, sales volume, and demand forecast; S2. Use data to build models, monitor inventory status and optimize space through models; S3. Conduct inventory counts to check whether the actual inventory matches the recorded inventory, and evaluate the actual quality and expiration date of the inventory; S4. Optimize demand forecasts, predict the demand and sales trends of pharmaceutical products, reduce inventory fluctuations, improve inventory turnover, and optimize inventory management by optimizing demand forecasts.

2. The pharmaceutical cold chain inventory optimization method based on neural network according to claim 1 is characterized in that: Step S1 is specifically as follows: S11. Identify the types of data that need to be collected, develop a detailed data collection plan, and collect data for the past 12 months based on the data types; S12. Extract the required data from various data sources, ensure the integrity and accuracy of the data, perform preliminary cleaning and verification of the data, and eliminate erroneous and inconsistent data; S13. Store the collected data in a data warehouse and establish a data management mechanism to ensure data accessibility, security and privacy protection; S14. Further clean and process the collected data, including removing duplicate data, filling missing values, and standardizing data formats; S15. Integrate data from different sources to form a unified data set and back up the data set.

3. The pharmaceutical cold chain inventory optimization method based on neural network according to claim 1 is characterized in that: Step S2 is specifically as follows: S21. According to business requirements and data characteristics, select appropriate features as the basis for clustering, and select clustering algorithm modeling according to data characteristics and clustering goals; S22. Use the selected clustering algorithm to perform cluster analysis on the data, classify similar data points into the same category, and analyze the clustering results to identify the characteristics and differences of different categories; S23. Develop targeted inventory management strategies based on the clustering results. For products with high inventory levels but low turnover rates, reduce purchases or adjust sales strategies. For products with large sales fluctuations, strengthen demand forecasting and inventory control.

4. The neural network-based pharmaceutical cold chain inventory optimization method according to claim 3, characterized in that: The clustering algorithm is selected according to the characteristics of the data and the clustering target to model the data. Specifically, the data set is divided into k clusters, each cluster is represented by a centroid, and the square error of the sample points in the cluster is minimized by iteratively updating the position of the centroid.

5. The pharmaceutical cold chain inventory optimization method based on neural network according to claim 1, characterized in that: Step S3 is specifically as follows: S31. Determine the time range for inventory taking and notify warehouse management personnel in advance; S32. Start the inventory at the scheduled time, ensure that all personnel are in place, scan the inventory items one by one, and record the actual inventory quantity; Compare the actual inventory quantity with the inventory quantity recorded in the system. For any mismatch, make detailed records and find out the reasons. S33. Check the appearance, packaging and labels of inventory goods to ensure that they are not damaged or expired.

6. The pharmaceutical cold chain inventory optimization method based on neural network according to claim 1, characterized in that: In step S4, the model is used to optimize the demand forecast, and the model expression is: y=β0+β1x1+β2x2+…+β p x p +∈; In the formula, y is the dependent variable, i.e., demand and sales trend, β0 is the intercept term, β1, β2, .., β p are regression coefficients, x1,x2,...,x p is the independent variable and ε is the error term.

7. The neural network-based pharmaceutical cold chain inventory optimization method according to claim 1, characterized in that: The optimization of inventory management in step S4 also needs to be performed in combination with a weighted algorithm, specifically: 1) Integrate the shelf life, distance to the destination and product type information of related products; 2) Convert each factor into a numerical value. The shelf life is calculated based on the remaining days, the distance to the destination is directly calculated using kilometers, and the product type is assigned different scores based on market demand and value. 3) Assign appropriate weight coefficients to each factor based on business needs and strategies; 4) Calculate the weighted score of each product through weighted calculation, sort all products according to the calculated weighted score, and give priority to products with high scores.

8. The neural network-based pharmaceutical cold chain inventory optimization method according to claim 7, characterized in that: After weighted calculation, if the same type of products have the same score, they will be shipped out on a first-in, first-out basis. Different types of products will have their final weighted score multiplied by a constant term, which is the score assigned to the product type based on market demand and value.

9. The neural network-based pharmaceutical cold chain inventory optimization method according to claim 8, characterized in that: The weighted score calculation formula for each product is calculated by weighted calculation: Weighted Score=w SL ×SL+w DD ×DD+w PT ×PT; In the formula, SL is the shelf life of the product, which is obtained by quantifying the remaining days; DD is the distance to the destination, which is obtained by quantifying the distance in kilometers; PT is the weight of the product type, with a score range of 1-5; The quantitative expression of the product's shelf life is: The quantitative expression of the distance in kilometers to the destination is: DD = actual kilometres received / quantified benchmark kilometres; Among them, the calculation formula for the quantitative benchmark kilometers is: Quantitative benchmark mileage = (actual mileage received × 2) × 0.618; that is, the total round-trip mileage to the delivery location multiplied by the golden ratio.

10. A pharmaceutical cold chain inventory optimization system based on a neural network, the system being used to implement the method of any one of claims 1 to 9, comprising: The data collection and preprocessing module is used to collect and preprocess the collected inventory data; The model building and analysis module is used to build models based on the data preprocessed by the data collection and preprocessing module, and formulate targeted inventory management strategies based on the model clustering results; The inventory counting and quality assessment module is used to assess the actual quality and validity period of the inventory based on the data preprocessed by the data collection and preprocessing module; The demand forecasting and inventory optimization module is used to optimize demand forecasting and reduce inventory fluctuations by optimizing demand forecasting, improve inventory turnover, and optimize inventory management.

Images