A hospital logistics procurement inventory management method and system based on artificial intelligence

Abstract

The present invention discloses an artificial intelligence-based hospital logistics procurement inventory management method and system, which relates to the technical field of hospital logistics and supply chain management. The method and system include: constructing a CIVS matrix, calculating dynamic EOQ, adjusting EOQ based on the CIVS matrix, calculating order cycle, and setting hospital logistics procurement safety stock; artificial intelligence provides optimal suppliers to hospital managers through multi-criteria decision analysis; constructing a MaMPA multi-objective optimization procurement plan, introducing low-carbon and environmental protection constraints, initializing prey populations, setting three iterative stages, randomly introducing FADs effects after each stage, calculating non-dominated sorting and crowding mechanism in the iterative process to determine predators, and using fuzzy decision making to select the optimal plan; artificial intelligence executes the entire process of automatic replenishment and purchase order generation of hospital logistics procurement inventory management, enhances the flexibility and adaptability of the supply chain by constructing the CIVS matrix, introduces low-carbon and environmental protection constraints by constructing the MaMPA multi-objective optimization procurement plan, and improves the scientific nature and optimization degree of procurement decisions.

Description

Technical Field

[0001] The present invention relates to the technical field of hospital logistics and supply chain management, and in particular to an artificial intelligence-based hospital logistics procurement inventory management method and system. Background Art

[0002] With the continuous expansion of hospital scale and the rapid growth of medical needs, hospital logistics procurement and inventory management have become increasingly complex. With the rapid development of technologies such as artificial intelligence (AI), big data, and machine learning, more and more hospitals have begun to explore intelligent management solutions. The introduction of artificial intelligence enables medical supplies procurement and inventory management to be data-driven, using historical data for accurate demand forecasting and dynamic optimization. It can effectively reduce excessive inventory backlogs, while reducing the risk of out-of-stocks and improving the overall efficiency of hospital supplies procurement. In addition, the combination of multi-criteria decision analysis (MCDA) and optimization algorithms can comprehensively consider multiple factors such as suppliers' delivery capabilities, prices, and quality, thereby providing hospital managers with the best supplier selection and ensuring the stability and economy of material supply.

[0003] However, the existing artificial intelligence-driven hospital logistics procurement inventory management technology still has some shortcomings in practical applications. It lacks comprehensive consideration of multiple influencing factors. In the setting of safety stocks and the calculation of order cycles, the existing multi-objective optimization algorithms often lack effective low-carbon and environmental protection constraints in actual applications, and fail to achieve a balance between environmental protection and economic benefits. Many systems have intelligent replenishment functions, but still lack full-process automated monitoring and dynamic adjustment mechanisms, resulting in hospitals being unable to respond to changes in inventory levels and supply chain conditions in real time during the management process. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an artificial intelligence-based hospital logistics procurement inventory management method to solve the problem of lack of comprehensive consideration of multiple influencing factors. In the setting of safety stock and calculation of ordering cycle, the existing multi-objective optimization algorithms often lack effective low-carbon and environmental protection constraints in actual application, and fail to achieve a balance between environmental protection and economic benefits. Many systems have intelligent replenishment functions, but still lack full-process automated monitoring and dynamic adjustment mechanisms, resulting in the hospital being unable to respond to changes in inventory levels and supply chain conditions in real time during the management process.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a hospital logistics procurement inventory management method based on artificial intelligence, which comprises:

[0008] Collect and preprocess medical supplies data to construct the CIVS matrix;

[0009] Build ARIMA models, calculate dynamic EOQ, adjust EOQ based on CIVS matrix, calculate order cycle, and set safety stock for hospital logistics procurement;

[0010] Based on the CIVS matrix, AI provides hospital managers with the best suppliers through multi-criteria decision analysis;

[0011] A multi-objective optimization procurement scheme for MaMPA was constructed, which introduced low-carbon and environmental protection constraints, initialized the prey population, set three iterative stages, and randomly introduced FADs effects after each stage. The iterative process calculated the non-dominated sorting and crowding mechanism to determine the predators, and used fuzzy decision making to select the optimal scheme.

[0012] Artificial intelligence executes the entire process of hospital logistics procurement inventory management, automatic replenishment and purchase order generation;

[0013] Automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management.

[0014] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the collection and preprocessing of medical supplies data and the construction of the CIVS matrix include:

[0015] The medical supplies data includes the structural data of all medicines, consumables and equipment in the hospital warehouse, pharmacy and using departments. The weighted cost score of each medical supply after preprocessing is calculated using the weighted average method, and the C classification of each medical supply is obtained using K-MeansWCSS;

[0016] Define three evaluation indicators, assign discrete values ​​to the three evaluation indicators of each medical material after preprocessing, and calculate 、 and , a three-layer decision tree was designed to classify medical supplies into I1, I2, and I3 according to the combination rules of E1, E2, and E3, and the importance scores of I1, I2, and I3 were set using the multiple method to obtain the I classification of each medical supply;

[0017] Extract the time-unit consumption series of each medical material from the preprocessed medical material data, calculate the demand volatility score of each medical material, and use hierarchical clustering based on Euclidean distance and Ward method to obtain the V classification of each medical material;

[0018] Obtain the life cycle environmental data of each medical supply, the reverse normalized carbon emissions and waste generation data, use the equal weighted average formula to calculate the sustainability score of each medical supply, and use K-MeansWCSS to obtain the S classification of each medical supply;

[0019] Based on the weighted cost score, importance score, demand volatility score and sustainability score of each medical supply as the quantitative scores of C, I, V and S indicators respectively;

[0020] Construct an evaluation matrix, normalize it, calculate the entropy value for each indicator, calculate the objective weight of each indicator, and calculate the weighted score for each medical supply;

[0021] Define I and S as positive indicators, C and V as negative indicators, and calculate the ideal solution for positive and negative indicators. and negative ideal solutions , calculate the Euclidean distance of each medical supply to the ideal solution and the negative ideal solution, calculate the relative proximity as the priority score of each medical supply, use quartile division to divide the priority order of each medical supply, and construct the CIVS matrix.

[0022] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the steps of constructing an ARIMA model, calculating dynamic EOQ, adjusting EOQ based on the CIVS matrix, calculating the order cycle, and setting the hospital logistics procurement safety stock include:

[0023] Construct an autoregressive integrated moving average (ARIMA) model, collect historical hospital inventory medical supply data, preprocess it, and perform training. Use the preprocessed medical supply data to predict the monthly demand for each medical supply and calculate the annual demand D.

[0024] Use the classic economic order quantity model EOQ to calculate the optimal order quantity for each medical supply , artificial intelligence obtains the priority weight of the CIVS indicator from the hospital logistics management system and the manager's input through the API, assigns a priority value to each subcategory of CIVS based on the category order, calculates the weighted priority index of each medical supply, and linearly maps the weighted priority index of each medical supply. Based on the CIVS matrix, artificial intelligence adjusts the optimal order quantity, calculates the order cycle, calculates the daily order cycle, calculates the adaptive adjustment coefficient, and calculates the adjusted daily order cycle;

[0025] Calculate the shortage cost of each medical supply, calculate the service level ratio, use the inverse cumulative distribution function of the standard normal distribution to map the service level ratio to the standard normal distribution, and calculate the hospital logistics procurement safety stock. , based on the quantitative score of V indicator, calculate and adjust the hospital logistics procurement safety stock , calculate the maximum hospital logistics procurement safety stock, set and adjust the hospital logistics procurement safety stock Set the inventory priority adjustment factor for the smallest hospital logistics procurement safety stock and adjust the largest hospital logistics procurement safety stock.

[0026] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the artificial intelligence provides the optimal supplier to the hospital manager through multi-criteria decision analysis, including:

[0027] Construct an AHP hierarchical structure, define the target layer as the decision goal of the optimal medical supplies supplier, the criterion layer as the criteria C, I, V and S data, and the solution layer as the alternative suppliers;

[0028] Based on the hospital's historical procurement data and priorities, AI generates importance scores for C, I, V, and S according to Saaty's 1-9 scale, constructs an initial pairwise comparison matrix, and presents the initial pairwise comparison matrix A to managers in tabular form. Managers can adjust the size of the matrix elements through the interface based on the hospital's actual needs, otherwise it remains unchanged.

[0029] Use the geometric mean method to calculate the geometric mean of the elements in each row of the adjusted pairwise comparison matrix as the preliminary weight of the criterion, and normalize it to obtain the final weight of the criterion. Generate a weight vector, calculate the weighted sum vector, calculate the maximum eigenvalue, and perform consistency detection. Based on the test results output and adjustment, use the weighted sum formula to calculate the comprehensive score for each supplier, and arrange them in descending order. The manager confirms the procurement decision and outputs the optimal supplier.

[0030] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the MaMPA multi-objective optimization procurement solution is constructed, three iterative stages are set, and fuzzy decision-making is used to select the optimal solution, including:

[0031] Based on the CIVS matrix, the MaMPA method is used to construct a multi-objective optimization objective function, and low-carbon and environmental protection constraints are introduced;

[0032] Define the procurement ratio, route, and vehicle as optimization objectives, mark them as prey, and use a random generator to generate the initial prey population;

[0033] The design consists of phase 1 where the prey position is updated using Brownian motion, phase 2 where the predator refines the optimal solution and the prey explores new paths, and phase 3 where local optimization is performed using Lévy motion.

[0034] Use the empirical method to set the maximum number of iterations, calculate the non-dominated sorting and crowding distance for the initial prey population, screen the optimal predator matrix of the current iteration, and recalculate the non-dominated sorting and crowding distance based on the updated prey after each subsequent iteration, update the optimal predator matrix, and execute stage 1 when the current number of iterations is less than one-third of the maximum number of iterations. When the number of iterations is greater than or equal to one-third of the maximum number of iterations and less than or equal to two-thirds of the maximum number of iterations, execute stage 2. When the number of iterations is greater than or equal to two-thirds of the maximum number of iterations, execute stage 3. Trigger the FADs effect after each execution. After reaching the maximum number of iterations, stop the iteration, output the final updated prey, calculate the non-dominated sorting and crowding distance, select the prey ranked in the top 50% of the crowding distance, and generate a priority solution subset.

[0035] Use fuzzy membership to select the optimal compromise solution from the priority solution subset, sort all the comprehensive memberships in descending order, and select the prey corresponding to the largest comprehensive membership as the optimal procurement ratio, route and vehicle.

[0036] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the artificial intelligence executes the entire process of automatic replenishment and purchase order generation for hospital logistics procurement inventory management, including:

[0037] Every adjustment day of the ordering cycle, artificial intelligence uses IoT sensors and inventory management systems to track the current inventory of each medical supply in real time, sends replenishment instructions based on the priority order of each medical supply in the CIVS matrix, generates electronic purchase orders using standard templates, and pushes the orders to the procurement department and suppliers via email through the intelligent supply chain management platform.

[0038] As a preferred solution of the artificial intelligence-based hospital logistics procurement inventory management method of the present invention, the entire process of automated monitoring and dynamic adjustment of hospital logistics procurement inventory management includes:

[0039] Artificial intelligence monitors the operation status of the entire process of hospital logistics procurement and inventory management in real time through the ERP system, uses SimCLR to detect abnormal operation status, and stores the entire process and abnormality logs in the hospital database;

[0040] The artificial intelligence performs self-inspections every week, checks external inputs, uses meta-learning MAML to evaluate system adaptability, and sets a task loss threshold. When the value of the system adaptability evaluated by meta-learning MAML is greater than the task loss threshold, the entire process of hospital logistics procurement and inventory management is updated.

[0041] In a second aspect, the present invention provides a hospital logistics procurement inventory management system based on artificial intelligence, comprising:

[0042] The collection and construction module is used to collect and preprocess medical supplies data and build the CIVS matrix;

[0043] The calculation setting module is used to build the ARIMA model, calculate the dynamic EOQ, adjust the EOQ based on the CIVS matrix, calculate the order cycle, and set the hospital logistics procurement safety stock;

[0044] The decision analysis module is used to provide hospital managers with the best suppliers through multi-criteria decision analysis based on the CIVS matrix.

[0045] The multi-objective optimization procurement module is used to construct the MaMPA multi-objective optimization procurement plan, introduce low-carbon and environmental protection constraints, initialize the prey population, set three iterative stages, randomly introduce FADs effects after each stage, iterate and calculate the non-dominated sorting and crowding mechanism to determine the predators, and use fuzzy decision making to select the optimal plan;

[0046] The replenishment generation module is used by artificial intelligence to execute the entire process of automatic replenishment and purchase order generation for hospital logistics and procurement inventory management;

[0047] The monitoring and adjustment module is used to automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management.

[0048] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the artificial intelligence-based hospital logistics procurement inventory management method as described in the first aspect of the present invention is implemented.

[0049] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the artificial intelligence-based hospital logistics procurement inventory management method as described in the first aspect of the present invention.

[0050] The beneficial effects of the present invention are as follows: the present invention collects and pre-processes medical supplies data to construct a CIVS matrix; constructs an ARIMA model, calculates dynamic EOQ, adjusts EOQ based on the CIVS matrix, calculates the order cycle, and sets a hospital logistics procurement safety stock; based on the CIVS matrix, artificial intelligence provides hospital managers with the best supplier through multi-criteria decision analysis; constructs a MaMPA multi-objective optimization procurement plan, introduces low-carbon and environmental protection constraints, initializes the prey population, sets three iterative stages, randomly introduces the FADs effect after each stage, and calculates the non-dominated sorting and crowding mechanism in the iterative process to determine the hunter, and uses fuzzy decision-making to select the optimal plan; artificial intelligence executes the entire process of automatic replenishment and purchase order generation for hospital logistics procurement inventory management, enhances the comprehensiveness and optimization of supply chain decisions, improves the flexibility and response speed of inventory management, optimizes the overall efficiency of procurement and inventory management, and enhances sustainability and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 This is a flowchart of the hospital logistics procurement inventory management method based on artificial intelligence in Example 1.

[0053] Figure 2 This is a schematic diagram of the hospital logistics procurement inventory management system based on artificial intelligence in Example 1.

[0054] Figure 3 Flowchart of the CIVS matrix construction process in Example 1.

[0055] Figure 4 Flowchart of the MaMPA optimization algorithm process in Example 1. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0059] Example 1, with reference to Figures 1 to 4 , which is the first embodiment of the present invention, provides a hospital logistics procurement inventory management method based on artificial intelligence, comprising the following steps:

[0060] S1, collect and preprocess medical supplies data and construct the CIVS matrix;

[0061] The medical supplies data includes the structural data of all medicines, consumables and equipment in the hospital warehouse, pharmacy and using departments;

[0062] The structural data of the drugs, consumables and equipment, including unit procurement cost, type, supplier delivery cycle, environmental impact factors, unit time consumption and real-time inventory;

[0063] The preprocessing includes using a Z-score anomaly detection algorithm to check and delete outliers in the medical supplies data, using a linear interpolation method to supplement missing values, and performing standardization;

[0064] Calculate the cost and economic index C of medical supplies. Use the weighted average method to calculate the weighted cost score of each medical supply after preprocessing. Set the number of clusters to 3 (corresponding to C1: high cost, C2: medium cost, and C3: low cost). Use K-MeansWCSS to minimize the intra-cluster sum of squares and assign the weighted cost score of each medical supply to the nearest cluster center using Euclidean distance to obtain the C classification of each medical supply (C1, C2, and C3).

[0065] To calculate the importance index I of medical supplies, we defined three evaluation indicators: clinical use (whether the medical supplies are used in key scenarios such as emergency treatment and surgery), consumption (the average annual consumption of medical supplies, reflecting the frequency of use), and substitutability (whether there are alternative medical supplies, which affects the necessity). We assigned discrete values ​​to the three evaluation indicators for each pre-processed medical supply. The formula is as follows:

[0066] ,

[0067] ,

[0068] ,

[0069] in, For the The urgency value of the clinical use of the medical supplies (1: urgent, 0: non-urgent), For the The consumption level of each medical material (1: high, 0: low), For the The substitutability value of the medical supplies (1: no substitute, 0: there is a substitute), For the The use of medical supplies, Medical supplies for emergency, surgical and intensive care use, For the Annual consumption of medical supplies is the total annual consumption of all medical supplies, is the median percentage of all medical supplies consumption. For the Number of alternatives to medical supplies;

[0070] Artificial intelligence extracts through text parsing (keyword matching) , extracted from the pre-processed medical supplies data and ,calculate 、 and ;

[0071] Design a three-layer decision tree with E1 as the root node, giving priority to judgment, as clinical urgency is the core indicator of medical importance. The second layer is E2, where high consumption further distinguishes importance. The third layer is E3, where medical supplies with no substitutes have higher priority.

[0072] According to the combination rules of E1, E2 and E3, medical supplies are classified into I1 (critical), I2 (essential) and I3 (replaceable), and the importance scores of I1, I2 and I3 are set using the multiple method, ranging from ;

[0073] The combination rule is as follows:

[0074] If :

[0075] If : I1

[0076] If :

[0077] If : I1

[0078] If : I2;

[0079] If :

[0080] If :

[0081] If : I2

[0082] If : I3

[0083] If : I3;

[0084] Extract the unit time consumption series of each medical material from the preprocessed medical material data and calculate the demand volatility score of each medical material. The formula is:

[0085] ,

[0086] in, For the Demand volatility scores for medical supplies For the The standard deviation of the monthly consumption of medical supplies, For the The average monthly consumption of various medical supplies;

[0087] Calculate the volatility index V of medical supply demand, set the target number of categories to 3 (V1: high volatility, V2: medium volatility, V3: low volatility), assign the volatility scores of all medical supplies to the V classification of each medical supply based on Euclidean distance and Ward method using hierarchical clustering;

[0088] To calculate the sustainability indicator S for medical supplies, AI sends electronic data requests to suppliers through the hospital's supply chain management system, SAP, to obtain lifecycle environmental data for each medical supply. For medical supplies for which the supplier does not provide data, the AI ​​queries the third-party LCA database, Ecoinvent, using the medical supply name as a keyword to obtain the average lifecycle environmental data for that medical supply.

[0089] The life cycle environmental data, including carbon emissions, recycling rates and waste generation data;

[0090] The normalized values ​​of the life cycle environmental data of each substance are calculated, and the carbon emissions and waste generation data after reverse normalization (to ensure that low carbon emissions and low waste generation correspond to high sustainability) are calculated using an equal-weighted weighted average formula to calculate the sustainability score of each medical supply. The formula is as follows:

[0091] ,

[0092] in, For the The sustainability score of each medical supply ranges from [0,1], where a higher score indicates stronger sustainability. For the The normalized reverse carbon emission value of medical supplies, For the The normalized recyclability value of medical supplies is For the Normalized reverse waste generation value of medical supplies;

[0093] Set the number of clusters to 3 (corresponding to S1: high sustainability, S2: medium sustainability, and S3: low sustainability). Use K-MeansWCSS to assign the sustainability scores of all medical supplies to the nearest cluster center using Euclidean distance by minimizing the intra-cluster sum of squares. This yields the S classification (S1, S2, and S3) and corresponding cluster center value for each medical supply.

[0094] Based on the weighted cost score, importance score, demand volatility score and sustainability score of each medical supply as the quantitative scores of C, I, V and S indicators respectively;

[0095] Each medical supply is used as a column of the matrix, and the quantitative scores of the C, I, V, and S indicators corresponding to each medical supply are used as rows of the matrix to construct an evaluation matrix and normalize it;

[0096] Based on the evaluation matrix, the entropy value is calculated for each indicator. The formula is as follows:

[0097] ,

[0098] in, For the The entropy value of the indicator, is the normalization constant, ensuring exist Inside, For the Medical supplies in the The ratio of indicators;

[0099] Based on the entropy value, the objective weight of each indicator is calculated as follows:

[0100] ,

[0101] in, For the The objective weight of each indicator;

[0102] Use the objective weight of each indicator to weight the indicator corresponding to each medical supply in the normalized evaluation matrix to obtain a weighted score for each medical supply;

[0103] Define I and S as positive indicators (the higher the better), and C and V as negative indicators (the lower the better);

[0104] Calculate the ideal solution for positive and negative indicators and negative ideal solutions , the formula is as follows:

[0105] ,

[0106] ,

[0107] in, For the Medical supplies in the The weighted score of the indicators, and Respectively The weighted scores of the indicators in the ideal solution and the negative ideal solution;

[0108] Calculate the Euclidean distance of each medical supply to the ideal solution and the negative ideal solution, use the Euclidean distance of each medical supply to the ideal solution and the negative ideal solution to calculate the relative proximity, which is used as the priority score of each medical supply. Use quartile division to divide the priority score of each medical supply into high priority, medium priority and low priority order. The C, I, V and S indicators, priority order and priority score of each medical supply form the rows of the CIVS matrix, and the type of each medical supply is used as the column of the CIVS matrix to construct the CIVS matrix.

[0109] The calculation of weighted cost scores and K-Means clustering can help hospitals make effective decisions on budget control and procurement strategies. The definition of evaluation indicators and the assignment of discrete values ​​provide quantitative input for intelligent decision-making, effectively reflecting the priority and urgency of medical supplies in different scenarios and providing a more scientific basis for subsequent decision tree design. By designing a three-layer decision tree with multi-level classification rules, the priority of supplies is more clearly defined, avoiding the waste or shortage of resources that may occur under a single standard. By calculating the demand volatility score for each medical supply and using hierarchical clustering to classify the supplies into three categories, hospitals can more flexibly formulate inventory strategies, ensuring uninterrupted supply of high-volatility supplies while avoiding inventory backlogs of low-volatility supplies. By obtaining lifecycle environmental data for medical supplies from suppliers or third-party databases and calculating the sustainability score for each supply, hospitals can not only reduce negative environmental impacts but also enhance their social responsibility image in environmental protection, in line with modern sustainable development concepts. The construction of the CIVS matrix and the prioritization provide a comprehensive and integrated evaluation for each supply, ensuring that hospitals consider inventory management and procurement decisions from a holistic perspective.

[0110] S2. Build an ARIMA model, calculate dynamic EOQ, adjust EOQ based on the CIVS matrix, calculate the order cycle, and set the hospital logistics procurement safety stock;

[0111] Specifically, historical hospital inventory data on medical supplies was collected and pre-processed;

[0112] Construct an autoregressive integrated moving average (ARIMA) model, taking preprocessed historical hospital inventory medical supplies data as input, using maximum likelihood estimation (MLE) as the optimization objective, and iterating through the numerical optimization algorithm BFGS to output the trained ARIMA model.

[0113] Input the preprocessed medical supplies data into the trained ARIMA model to predict the monthly demand for each medical supply, and multiply the monthly demand by 12 to obtain the annual demand D.

[0114] AI aggregates the fixed cost W of recent orders and the purchase price of each medical supply from procurement records via API. , obtain the annual holding cost H of medical supplies per unit from the inventory management system;

[0115] Use the classic economic order quantity model EOQ to calculate the optimal order quantity for each medical supply ;

[0116] Calculating Basis Adjustment Factors , the formula is as follows:

[0117] ,

[0118] in, is the average annual demand;

[0119] Calculate the mapping lower limit a using the following formula:

[0120] ,

[0121] in, The maximum EOQ adjustment range is obtained from the hospital logistics management system and managers. is the standardized standard deviation of the annual demand D;

[0122] Calculate the mapping upper limit b, the formula is as follows:

[0123] ,

[0124] The AI ​​obtains the priority weights of CIVS indicators from the hospital logistics management system and managers’ inputs through the API. If not provided, the priority weights are assigned equally by default;

[0125] For each subcategory of CIVS, priority values ​​were assigned based on the order of the categories: C1–C3, I1–I3, V1–V3, and S1–S3, all assigned 1–3 points in sequence;

[0126] Calculate the weighted priority index for each medical supply using the following formula:

[0127] ,

[0128] in, For the Weighted priority index of medical supplies, 、 、 and Respectively The priority weights of various medical supplies in each indicator, 、 、 and Respectively The priority values ​​of various medical supplies in each indicator;

[0129] Use the linear mapping formula to map the weighted priority index of each medical supply to the upper and lower limits. The formula is as follows:

[0130] ,

[0131] in, is the CIVS adjustment factor;

[0132] Artificial intelligence adjusts the optimal order quantity based on the CIVS matrix , the formula is as follows:

[0133] ,

[0134] in, For the The adjusted optimal order quantity of the medical supplies;

[0135] Divide the adjusted optimal order quantity by the monthly forecast demand to get the order cycle, and multiply it by the number of days in the month to get the daily order cycle;

[0136] Based on the quantitative scores of the I and V indicators, the adaptive adjustment coefficient is calculated using the following formula:

[0137] ,

[0138] in, is the order cycle adjustment coefficient, and They are the quantitative scores of I and V indicators, ranging from , As the reference coefficient, ensure Do not deviate excessively from reasonable ranges;

[0139] Order cycle adjustment factor Multiply the daily order cycle as the adjusted daily order cycle;

[0140] AI obtains the delivery lead time L (days) and historical shortage frequency (times / year) FS from suppliers;

[0141] Quantitative score based on I indicator , calculate the shortage cost of each medical supply, the formula is as follows:

[0142] ,

[0143] in, For the The cost of shortages of medical supplies, is the shortage cost amplification factor;

[0144] Based on calculating the shortage cost of each medical supply , calculate the service level ratio , the formula is as follows:

[0145] ,

[0146] The service level ratio is calculated using the inverse cumulative distribution function of the standard normal distribution. Mapped to the standard normal distribution, the service level coefficient is obtained ;

[0147] Based on service level factor , calculate hospital logistics procurement safety stock , the formula is as follows:

[0148] ,

[0149] in, is the annual demand standard deviation, To standardize lead time to monthly units;

[0150] Calculate and adjust hospital logistics procurement safety stock based on the quantitative score of V indicator , the formula is as follows:

[0151] ,

[0152] The hospital's logistics procurement safety stock will be adjusted and adjust the optimal order quantity Add them together as the maximum hospital logistics procurement safety stock, set and adjust the hospital logistics procurement safety stock Procurement of safety stock for the smallest hospital logistics;

[0153] Based on the classification of I and V indicators corresponding to the medical supply types (such as I1-V1), nine inventory priority adjustment factors are set to adjust the maximum hospital logistics procurement safety stock. The formula is as follows:

[0154] ,

[0155] in, To adjust the largest hospital logistics procurement safety stock, Procurement of safety stock for the largest hospital logistics, is the inventory priority adjustment factor, is the adjustment factor of the S indicator, which refers to the quantitative score of the standardized S indicator;

[0156] The setting of the nine inventory priority adjustment factors refers to standardizing the quantitative scores of the I and V indicators corresponding to the medical supplies and then performing weighted summation.

[0157] By using the EOQ model to calculate the optimal order quantity, the optimal order quantity for each medical supply can be accurately calculated. By assigning weighted priorities to various indicators based on the CIVS matrix, hospitals can reasonably adjust procurement decisions based on factors such as the actual demand, supply situation, and cost of each medical supply. By combining the quantitative scores of the I and V indicators, different types of medical supplies can be prioritized according to the actual needs of the hospital and the urgency of procurement, thereby optimizing inventory allocation and procurement cycles.

[0158] S3. Based on the CIVS matrix, AI provides hospital managers with the best suppliers through multi-criteria decision analysis;

[0159] Specifically, the quantitative scores of the C, I, V, and S indicators are normalized and used as the criterion C, I, V, and S data;

[0160] Construct an AHP hierarchical structure, define the target layer as the decision goal of the optimal medical supplies supplier, the criterion layer as the criteria C, I, V and S data, and the solution layer as the alternative suppliers;

[0161] Normalize the total cost (purchase price and transportation price) of medical supplies provided by each supplier, and normalize the normalized total cost on the standard data to obtain the standard cost of each supplier. , Guidelines , Guidelines and guidelines The normalized score of

[0162] Based on the hospital's historical procurement data and priorities (e.g., if green supply chain is prioritized, S is more important than C), AI generates importance scores for C, I, V, and S according to Saaty's 1-9 scale.

[0163] Construct the initial pairwise comparison matrix, the formula is as follows:

[0164] ,

[0165] Among them, A is the pairwise comparison matrix, As a guideline Standards The importance ratio of

[0166] Artificial intelligence presents the initial pairwise comparison matrix A to managers in tabular form through the interactive interface of the intelligent supply chain management platform. Managers can adjust the size of matrix elements through the interface based on the actual needs of the hospital (for example, I is more important during an epidemic), otherwise it remains unchanged.

[0167] The geometric mean method is used to calculate the geometric mean of each row element of the adjusted pairwise comparison matrix as the preliminary weight of the criterion, and then normalized to obtain the final weight of the criterion and generate a weight vector;

[0168] Multiply the pairwise comparison matrix of the manager by the weight vector to obtain the weighted sum vector and calculate the maximum eigenvalue , the formula is as follows:

[0169] ,

[0170] in, is the weighted sum vector elements, is the weight of the mth criterion;

[0171] Calculate the consistency index CI, the formula is as follows:

[0172] ,

[0173] in, is the number of criteria;

[0174] The consistency ratio CR is calculated as follows:

[0175] ,

[0176] in, is the random consistency index, when n=4, RI=0.9 (based on the AHP standard table);

[0177] like , then the consistency of the pairwise comparison matrix of the managers is acceptable, the final weight of the criterion is valid, and it is directly output;

[0178] like , AI prompts managers through the platform to readjust the comparison matrix;

[0179] For each supplier, a weighted summation formula is used to calculate the overall score, as follows:

[0180] ,

[0181] in, For suppliers The comprehensive score of 、 、 and The criteria are , Guidelines , Guidelines and guidelines The final weight of the criteria that passed the consistency test, 、 、 and Suppliers In the guidelines , Guidelines , Guidelines and guidelines The normalized score of

[0182] All suppliers are sorted in descending order based on their comprehensive scores. AI, through the intelligent supply chain management platform, selects the supplier with the highest score by default and recommends it to managers, along with a dynamic chart display. Managers can review the recommendations through the platform interface, confirm their purchasing decisions, or adjust the final weights of the criteria, recalculate, and output the optimal supplier.

[0183] The dynamic graph includes generating a bar graph by comparing the comprehensive scores of each supplier and generating a radar graph by comparing the scores of each supplier on C, I, V, and S.

[0184] By constructing an AHP hierarchical model consisting of a target layer (selecting the optimal supplier), a criterion layer (C, I, V, S indicators), and a solution layer (various alternative suppliers), this model achieves structured modeling of complex procurement decision-making problems, clarifies the logical relationship between evaluation factors, and uses artificial intelligence to automatically generate criterion importance scores based on the hospital's historical procurement data and current policy priorities (such as prioritizing the S indicator for green supply chains). A targeted initial judgment matrix is ​​constructed, reducing the cognitive burden on managers and improving the initial quality of empowerment, avoiding human experience errors. By multiplying each supplier's score under the four criteria with the final criterion weight and taking a weighted sum, subjective, guesswork-based evaluations are avoided and the overall performance of different suppliers can be scientifically compared. By sorting the calculated supplier comprehensive scores in descending order, the supplier with the highest score is recommended by default, and the score structure is visualized using bar charts and radar charts. Managers can further optimize the weight configuration based on the charts or directly confirm the decision, ultimately achieving an explainable, interventionable, and visual supplier intelligent selection process, significantly improving management efficiency and decision-making confidence.

[0185] S4. Construct a MaMPA multi-objective optimization procurement plan, introduce low-carbon and environmental protection constraints, initialize the prey population, set three iterative stages, randomly introduce FADs effects after each stage, calculate the non-dominated sorting and crowding mechanism to determine the predators through the iterative process, and use fuzzy decision making to select the optimal plan;

[0186] Specifically, based on the CIVS matrix, with the goals of minimizing total procurement costs, maximizing the importance of medical supplies, minimizing demand fluctuation risks, maximizing sustainability, and minimizing carbon emissions, the MaMPA method is used to construct a multi-objective optimization objective function. The formula is as follows:

[0187] ,

[0188] in, is a multi-objective function, which represents the optimized target vector and contains five objectives. is the comprehensive cost (yuan), including procurement and transportation costs, is the total number of all medical supplies, For the i-th medical supplies from supplier Unit purchase cost (yuan / piece), For the i-th medical supplies from supplier The proportion of purchases, is the fixed purchase quantity of the i-th medical supplies, which is obtained by adjusting the maximum hospital logistics purchase safety stock minus the current inventory. is the number of suppliers, is the total number of transport routes, is the total number of transport vehicles, For route Use of vehicle Unit transportation cost (yuan / km), For route Use of vehicle Transport distance (km), is the quantitative score of the S indicator, is logistics carbon emissions (kg CO2), For vehicles Unit emission factor (kgCO2 / km);

[0189] For low-carbon and environmental protection, constraints are introduced into the above multi-objective optimization objective function. The formula is as follows:

[0190] ,

[0191] in, For route The collection of medical supplies involved, is the weight of the i-th medical supply, For the The maximum weight of the vehicle, For route transportation time, The earliest and latest permitted times for transportation;

[0192] Define the proportion of purchases ,route and vehicles To optimize the target, mark it as prey , For the prey, using a random generator to generate Initial prey population;

[0193] Perform non-dominated sorting on the initial prey population and identify the Pareto frontier set (non-dominated solution set) using the following formula:

[0194] ,

[0195] ,

[0196] in, for and The dominant relationship between is the dominant symbol, and Respectively and prey, h is the index of the five objectives in the multi-objective optimization objective function, and Don't and The value on the target h, For the The non-dominated frontier is a set of solutions that contains all solutions ranked The non-dominated solutions of , which do not dominate each other in the current level, but are dominated by the solutions of higher levels, for The ranking level of

[0197] For each non-dominated solution Calculate the congestion distance using the following formula:

[0198] ,

[0199] in, For the The crowding distance of prey, and are respectively the target h and The values ​​of the neighboring solutions (sorted by target h), is the maximum and minimum value of the target h in the Pareto front set;

[0200] The solutions in the Pareto front set are sorted in descending order according to the crowding distance, and the top 50% of prey are retained to generate the optimal predator matrix;

[0201] The system is designed in three phases: Phase 1, which uses Brownian motion to update the prey's position and focuses on global search; Phase 2, which uses Lévy motion to focus on local optimization of prey and predator, where the predator refines the optimal solution and the prey explores new paths, balancing exploration and exploitation; and Phase 3, which uses Lévy motion to focus on local optimization of prey and predator.

[0202] In the first stage, the formula is as follows:

[0203] ,

[0204] ,

[0205] in, For the Phase 1 of the iteration The step vector of the prey represents the direction and magnitude of the prey position update in the search space, is a Brownian motion random vector, generated based on normal distribution, For the The first iteration The optimal predator matrix represents the current optimal supplier allocation, route and vehicle combination, is the element-wise multiplication operator, is a uniform random vector with a value in between, For the Phase 1 of the iteration A newer prey;

[0206] In the stage 2, the formula is as follows:

[0207] ,

[0208] ,

[0209] in, For the Phase 2 of the first iteration The step vector of the prey, For the Phase 2 of the first iteration A newer prey, It is an adaptive step size control, which decreases with iteration;

[0210] The formula for stage 3 is as follows:

[0211] ,

[0212] ,

[0213] in, For the Phase 3 of the iteration The step vector of the prey, is the Lévy motion random vector (obtained based on Lévy distribution), For the Phase 3 of the iteration A newer prey;

[0214] The FADs effect is introduced and triggered after each stage. The formula is as follows:

[0215] ,

[0216] in, The updated FADs effect Prey location, For the The first iteration prey positions (S=1, 2, 3 correspond to the outputs of stages 1, 2, and 3), and are variable ranges (allocation ratio, route, vehicle), is the impact factor, is a random number (0 to 1), and Respectively The randomly selected and The location of the prey, for prey location (including proportions, routes, and vehicles sourced);

[0217] Use the empirical method to set the maximum number of iterations, calculate the non-dominated sorting and crowding distance for the initial prey population, screen the optimal predator matrix of the current iteration, and recalculate the non-dominated sorting and crowding distance based on the updated prey after each subsequent iteration, update the optimal predator matrix, and execute stage 1 when the current number of iterations is less than one-third of the maximum number of iterations. When the number of iterations is greater than or equal to one-third of the maximum number of iterations and less than or equal to two-thirds of the maximum number of iterations, execute stage 2. When the number of iterations is greater than or equal to two-thirds of the maximum number of iterations, execute stage 3. Trigger the FADs effect after each execution. After reaching the maximum number of iterations, stop the iteration, output the final updated prey, calculate the non-dominated sorting and crowding distance, select the prey ranked in the top 50% of the crowding distance, and generate a priority solution subset.

[0218] Use fuzzy membership to select the optimal compromise solution from the priority solution subset to balance the five-dimensional objectives. The formula is as follows:

[0219] ,

[0220] l ,

[0221] in, for On target The membership degree on xc for On target The solution above, and Target The minimum and maximum values ​​in the updated subset of prioritized solutions, is the number of solutions in the updated priority solution subset, l for The comprehensive membership degree;

[0222] Arrange all comprehensive memberships in descending order, and select the prey corresponding to the largest comprehensive membership as the optimal procurement ratio, route, and vehicle.

[0223] By constructing a multi-objective optimization objective function based on the CIVS matrix, balancing different needs by comprehensively considering multi-dimensional objectives, and realizing refined management of procurement decisions, hospitals can not only focus on economic costs in the procurement process, but also incorporate long-term goals such as sustainability and environmental protection, thereby improving the scientific nature and social responsibility of hospital logistics procurement. By adopting the MaMPA method for multi-objective optimization, the non-dominated sorting and crowding distance calculation of prey populations, through complex multi-objective optimization, hospitals can achieve a balance of multiple goals, ensuring that both procurement costs can be reduced and the stability and sustainability of material supply can be guaranteed. The crowding calculation ensures the diversity of solutions, avoids falling into local optimal solutions, and improves the comprehensiveness of the optimization effect. The introduction of the layout The Lang motion and Lévy motion enable the search strategy to be flexibly adjusted during the optimization process. The FADs effect is introduced after each stage to trigger the random position update of the prey, avoid the trap of local optimal solutions, and improve the global search capability. After each iteration, the optimal solution is continuously screened out through non-dominated sorting and congestion calculation, and the solution is refined through multiple iterations. It can gradually approach the optimal solution in each round of optimization, avoiding the problem of unstable results caused by insufficient calculation accuracy or improper setting of stopping conditions in traditional optimization methods. Through the fuzzy membership method, the optimal compromise solution is selected, the pros and cons of each solution under different objectives are quantified, and the multi-objective optimization results are converted into an operational decision support tool to help hospital managers make the best choice among multiple candidate solutions.

[0224] S5. Artificial intelligence executes the entire process of hospital logistics procurement inventory management, automatic replenishment, and purchase order generation;

[0225] Specifically, every adjustment day of the ordering cycle, artificial intelligence uses IoT sensors and the inventory management system to track the current inventory of each medical supply in real time. If the current inventory is less than or equal to the minimum hospital logistics procurement safety stock, a replenishment instruction is sent based on the priority order of each medical supply in the CIVS matrix, and an electronic purchase order is generated using a standard template (medical supply type, purchase quantity (adjusted maximum hospital logistics procurement safety stock minus current inventory), total price (purchase price and transportation cost), optimal supplier, purchase ratio, route, vehicle and status (such as "confirmed", "shipping", "delivered")). The order is pushed to the procurement department (for approval and record keeping) and the supplier (for confirmation and delivery arrangement) via email through the intelligent supply chain management platform.

[0226] By adjusting the ordering cycle and conducting inventory management based on real-time data, the accuracy and flexibility of inventory control have been improved. Through precise control of safety stock settings, timely replenishment of medical supplies has been achieved and the risk of out-of-stocks has been reduced. By optimizing the replenishment sequence based on the CIVS matrix, the scientific nature and priority sorting of procurement decisions have been improved. Through the standardized electronic purchase order generation and push system, the efficiency and transparency of the procurement process have been improved. Through email push on the intelligent supply chain management platform, the information flow and response speed between procurement and suppliers have been improved.

[0227] S6. Automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management.

[0228] Specifically, AI monitors the operating status of the entire process of hospital logistics procurement and inventory management (data collection, CIVS classification, demand forecasting, inventory optimization, and procurement execution) in real time through the ERP system, uses SimCLR to detect operational anomalies, and stores the entire process and anomaly logs in the hospital database;

[0229] The AI ​​conducts weekly self-inspections, checks external inputs (new medical supplies, new suppliers, and new requirements, such as increased weighting for green procurement), uses meta-learning MAML to evaluate system adaptability, and sets a task loss threshold. When the value of the system adaptability evaluated by meta-learning MAML is greater than the task loss threshold, the entire process of hospital logistics procurement and inventory management is updated.

[0230] Through artificial intelligence, the entire process of hospital logistics procurement and inventory management is monitored in real time, reducing manual intervention and operational errors, allowing each link to collaborate efficiently. By using SimCLR to detect abnormal operating status, anomalies and potential problems can be discovered in a timely manner. Through weekly self-inspections and checks on external inputs, it is ensured that the hospital logistics procurement and inventory management system can quickly adapt to the market environment and external changes. Through meta-learning MAML, the task loss threshold is set to ensure that the entire process will only be updated when the system's adaptability value exceeds the preset threshold. Rapid learning from a small amount of sample data enables the system to quickly adapt to and optimize its own operating strategies when faced with new procurement needs or changes, allowing the hospital to continuously optimize the procurement management process, adapt to the ever-changing medical supplies market, and improve the intelligence level of hospital logistics management.

[0231] This embodiment also provides an artificial intelligence-based hospital logistics procurement inventory management system, including:

[0232] The collection and construction module is used to collect and preprocess medical supplies data and build the CIVS matrix;

[0233] The calculation setting module is used to build the ARIMA model, calculate the dynamic EOQ, adjust the EOQ based on the CIVS matrix, calculate the order cycle, and set the hospital logistics procurement safety stock;

[0234] The decision analysis module is used to provide hospital managers with the best suppliers through multi-criteria decision analysis based on the CIVS matrix.

[0235] The multi-objective optimization procurement module is used to construct the MaMPA multi-objective optimization procurement plan, introduce low-carbon and environmental protection constraints, initialize the prey population, set three iterative stages, randomly introduce FADs effects after each stage, iterate and calculate the non-dominated sorting and crowding mechanism to determine the predators, and use fuzzy decision making to select the optimal plan;

[0236] The replenishment generation module is used by artificial intelligence to execute the entire process of automatic replenishment and purchase order generation for hospital logistics and procurement inventory management;

[0237] The monitoring and adjustment module is used to automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management.

[0238] This embodiment also provides a computer device suitable for the case of an artificial intelligence-based hospital logistics procurement inventory management method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the artificial intelligence-based hospital logistics procurement inventory management method proposed in the above embodiment.

[0239] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0240] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the artificial intelligence-based hospital logistics procurement inventory management method proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0241] In summary, the present invention collects and preprocesses medical supplies data to construct a CIVS matrix; constructs an ARIMA model to calculate dynamic EOQ, adjusts EOQ based on the CIVS matrix, calculates the order cycle, and sets a hospital logistics procurement safety stock; based on the CIVS matrix, artificial intelligence provides hospital managers with the best supplier through multi-criteria decision analysis; constructs a MaMPA multi-objective optimization procurement plan, introduces low-carbon and environmental protection constraints, initializes the prey population, sets three iterative stages, randomly introduces the FADs effect after each stage, and the iterative process calculates the non-dominated sorting and congestion mechanism to determine the hunter, and uses fuzzy decision-making to select the optimal plan; artificial intelligence executes the entire process of automatic replenishment and purchase order generation for hospital logistics procurement inventory management, thereby enhancing the comprehensiveness and optimization of supply chain decisions, improving the flexibility and response speed of inventory management, optimizing the overall efficiency of procurement and inventory management, and enhancing sustainability and environmental protection.

[0242] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A hospital logistics procurement inventory management method based on artificial intelligence, characterized by: include, Collect and preprocess medical supplies data to construct the CIVS matrix; Build an ARIMA model to predict the monthly demand for each medical supply. Use the classic economic order quantity model (EOQ) to calculate the optimal order quantity for each medical supply. Based on the CIVS matrix, adjust the optimal order quantity, calculate the order cycle, and set the hospital logistics procurement safety stock. Based on the CIVS matrix, an AHP hierarchy is constructed to provide hospital managers with the best suppliers; A multi-objective optimization procurement scheme for MaMPA was constructed, which introduced low-carbon and environmental protection constraints, initialized the prey population, set three iterative stages, and randomly introduced FADs effects after each stage. The iterative process calculated the non-dominated sorting and crowding distance to determine the predators, and used fuzzy decision making to select the optimal scheme. Artificial intelligence executes the entire process of hospital logistics procurement inventory management, automatic replenishment and purchase order generation; Automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management; The collection and preprocessing of medical supplies data to construct the CIVS matrix includes: The medical supplies data includes the structural data of all medicines, consumables and equipment in the hospital warehouse, pharmacy and using departments. The weighted cost score of each medical supply after preprocessing is calculated using the weighted average method, and the C classification of each medical supply is obtained using K-MeansWCSS; Define three evaluation indicators, assign discrete values ​​to the three evaluation indicators of each medical material after preprocessing, and calculate 、 and , To index the types of medical supplies, a three-layer decision tree was designed. According to the combination rules of E1, E2 and E3, medical supplies were classified into I1, I2 and I3. The importance scores of I1, I2 and I3 were set using the multiple method to obtain the I classification of each medical supply; Extract the time-unit consumption series of each medical material from the preprocessed medical material data, calculate the demand volatility score of each medical material, and use hierarchical clustering based on Euclidean distance and Ward method to obtain the V classification of each medical material; Obtain the life cycle environmental data of each medical supply, the reverse normalized carbon emissions and waste generation data, use the equal weighted average formula to calculate the sustainability score of each medical supply, and use K-MeansWCSS to obtain the S classification of each medical supply; Based on the weighted cost score, importance score, demand volatility score and sustainability score of each medical supply as the quantitative scores of C, I, V and S indicators respectively; Construct an evaluation matrix, normalize it, calculate the entropy value for each indicator, calculate the objective weight of each indicator, and calculate the weighted score for each medical supply; Define I and S as positive indicators, C and V as negative indicators, and calculate the ideal solution for positive and negative indicators. and negative ideal solutions , calculate the Euclidean distance of each medical supply to the ideal solution and the negative ideal solution, calculate the relative proximity as the priority score of each medical supply, use quartile division to divide the priority order of each medical supply, and construct the CIVS matrix.

2. The artificial intelligence-based hospital logistics procurement inventory management method according to claim 1, characterized in that: The method involves building an ARIMA model to predict the monthly demand for each medical supply, using the classic economic order quantity model (EOQ) to calculate the optimal order quantity for each medical supply, adjusting the optimal order quantity based on the CIVS matrix, calculating the order cycle, and setting the hospital logistics procurement safety stock, including: Construct an autoregressive integrated moving average (ARIMA) model, collect historical hospital inventory medical supply data, preprocess it, and perform training. Use the preprocessed medical supply data to predict the monthly demand for each medical supply and calculate the annual demand D. Use the classic economic order quantity model EOQ to calculate the optimal order quantity for each medical supply , artificial intelligence obtains the priority weight of the CIVS indicator from the hospital logistics management system and the manager's input through the API, assigns a priority value to each subcategory of the CIVS matrix based on the category order, calculates the weighted priority index of each medical supply, and linearly maps the weighted priority index of each medical supply. Based on the CIVS matrix, artificial intelligence adjusts the optimal order quantity, calculates the order cycle, calculates the daily order cycle, calculates the adaptive adjustment coefficient, and calculates the adjusted daily order cycle; Calculate the shortage cost of each medical supply, calculate the service level ratio, use the inverse cumulative distribution function of the standard normal distribution to map the service level ratio to the standard normal distribution, and calculate the hospital logistics procurement safety stock. , based on the quantitative score of V indicator, calculate and adjust the hospital logistics procurement safety stock , calculate the maximum hospital logistics procurement safety stock, set and adjust the hospital logistics procurement safety stock Set the inventory priority adjustment factor for the smallest hospital logistics procurement safety stock and adjust the largest hospital logistics procurement safety stock.

3. The artificial intelligence-based hospital logistics procurement inventory management method according to claim 2, characterized in that: The AHP hierarchy is constructed to provide hospital managers with the best suppliers, including: Construct an AHP hierarchical structure, define the target layer as the decision goal of the optimal medical supplies supplier, the criterion layer as the criteria C, I, V and S data, and the solution layer as the alternative suppliers; Based on the hospital's historical procurement data and priorities, AI generates importance scores for C, I, V, and S according to Saaty's 1-9 scale, constructs an initial pairwise comparison matrix, and presents the initial pairwise comparison matrix to managers in tabular form. Managers can adjust the size of the matrix elements through the interface based on the hospital's actual needs, otherwise it remains unchanged. Use the geometric mean method to calculate the geometric mean of the elements in each row of the adjusted pairwise comparison matrix as the preliminary weight of the criterion, and normalize it to obtain the final weight of the criterion. Generate a weight vector, calculate the weighted sum vector, calculate the maximum eigenvalue, and perform consistency detection. Based on the test results output and adjustment, use the weighted sum formula to calculate the comprehensive score for each supplier, and arrange them in descending order. The manager confirms the procurement decision and outputs the optimal supplier.

4. The artificial intelligence-based hospital logistics procurement inventory management method according to claim 3, characterized in that: The MaMPA multi-objective optimization procurement scheme is constructed, with three iterative stages and fuzzy decision making used to select the optimal scheme, including: Based on the CIVS matrix, the MaMPA method is used to construct a multi-objective optimization objective function, and low-carbon and environmental protection constraints are introduced; Define the procurement ratio, route, and vehicle as optimization objectives, mark them as prey, and use a random generator to generate the initial prey population; The design consists of phase 1 where the prey position is updated using Brownian motion, phase 2 where the predator refines the optimal solution and the prey explores new paths, and phase 3 where local optimization is performed using Lévy motion. Use the empirical method to set the maximum number of iterations, calculate the non-dominated sorting and crowding distance for the initial prey population, screen the optimal predator matrix of the current iteration, and recalculate the non-dominated sorting and crowding distance based on the updated prey after each subsequent iteration, update the optimal predator matrix, and execute stage 1 when the current number of iterations is less than one-third of the maximum number of iterations. When the number of iterations is greater than or equal to one-third of the maximum number of iterations and less than or equal to two-thirds of the maximum number of iterations, execute stage 2. When the number of iterations is greater than or equal to two-thirds of the maximum number of iterations, execute stage 3. Trigger the FADs effect after each execution. After reaching the maximum number of iterations, stop the iteration, output the final updated prey, calculate the non-dominated sorting and crowding distance, select the prey ranked in the top 50% of the crowding distance, and generate a priority solution subset. Use fuzzy membership to select the optimal compromise solution from the priority solution subset, sort all the comprehensive memberships in descending order, and select the prey corresponding to the largest comprehensive membership as the optimal procurement ratio, route and vehicle.

5. The artificial intelligence-based hospital logistics procurement inventory management method according to claim 4, characterized in that: The AI ​​executes the entire process of hospital logistics procurement inventory management, automatic replenishment, and purchase order generation, including: Every adjustment day of the ordering cycle, artificial intelligence uses IoT sensors and inventory management systems to track the current inventory of each medical supply in real time, sends replenishment instructions based on the priority order of each medical supply in the CIVS matrix, generates electronic purchase orders using standard templates, and pushes the orders to the procurement department and suppliers via email through the intelligent supply chain management platform.

6. The artificial intelligence-based hospital logistics procurement inventory management method according to claim 5, characterized in that: The entire process of automated monitoring and dynamic adjustment of hospital logistics procurement inventory management includes: Artificial intelligence monitors the operation status of the entire process of hospital logistics procurement and inventory management in real time through the ERP system, uses SimCLR to detect abnormal operation status, and stores the entire process and abnormality logs in the hospital database; The artificial intelligence performs self-inspections every week, checks external inputs, uses meta-learning MAML to evaluate system adaptability, and sets a task loss threshold. When the value of the system adaptability evaluated by meta-learning MAML is greater than the task loss threshold, the entire process of hospital logistics procurement and inventory management is updated.

7. An artificial intelligence-based hospital logistics procurement inventory management system, based on the artificial intelligence-based hospital logistics procurement inventory management method according to any one of claims 1 to 6, characterized in that: include, The collection and construction module is used to collect and preprocess medical supplies data and build the CIVS matrix; The calculation setting module is used to build the ARIMA model, calculate the dynamic EOQ, adjust the EOQ based on the CIVS matrix, calculate the order cycle, and set the hospital logistics procurement safety stock; The decision analysis module is used to provide hospital managers with the best suppliers through multi-criteria decision analysis based on the CIVS matrix. The multi-objective optimization procurement module is used to construct the MaMPA multi-objective optimization procurement plan, introduce low-carbon and environmental protection constraints, initialize the prey population, set three iterative stages, randomly introduce FADs effects after each stage, iteratively calculate the non-dominated sorting and crowding distance to determine the predators, and use fuzzy decision making to select the optimal plan; The replenishment generation module is used by artificial intelligence to execute the entire process of automatic replenishment and purchase order generation for hospital logistics and procurement inventory management; The monitoring and adjustment module is used to automatically monitor and dynamically adjust the entire process of hospital logistics procurement and inventory management.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the hospital logistics procurement inventory management method based on artificial intelligence according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the artificial intelligence-based hospital logistics procurement inventory management method according to any one of claims 1 to 6 are implemented.

Images