Artificial intelligence-based spd consumable demand prediction and intelligent replenishment method and system

By optimizing the SPD consumable demand prediction model through artificial intelligence-based data processing and genetic algorithms, the problems of low prediction accuracy and poor replenishment strategies in existing technologies that rely on experience have been solved. This has enabled accurate prediction of consumable demand and intelligent replenishment, thereby improving hospital operational efficiency and the stability of clinical diagnosis and treatment.

CN122266692APending Publication Date: 2026-06-23FENGHE (BEIJING) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610456019.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-08
Publication Date
2026-06-23

Smart Images

  • Figure CN122266692A_ABST
    Figure CN122266692A_ABST
Patent Text Reader

Abstract

This invention discloses an artificial intelligence-based method and system for predicting and intelligently replenishing SPD consumable demand, belonging to the field of medical consumable supply chain management technology. It obtains key influencing factors of SPD consumables by clustering analysis of historical SPD consumable demand correlation data and identifying influencing factors. Furthermore, through the acquisition and training of an initial SPD consumable demand prediction model, an accurate SPD consumable demand prediction model is obtained. This model is then used to predict real-time SPD consumable demand correlation data, outputting SPD consumable demand prediction results that better match actual needs and applications. In addition, the replenishment objective function is iteratively optimized using a multi-population genetic algorithm to obtain optimal SPD control parameters, thereby automating the generation of intelligent replenishment strategies.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of medical consumables supply chain management technology, specifically relating to an artificial intelligence-based method and system for predicting demand and intelligent replenishment of SPD consumables. Background Technology

[0002] With the rapid development of medical technology and the continuous deepening of medical services, the types, specifications and usage of medical consumables have all shown a significant upward trend, and their management has become a core aspect of hospital operating cost control and medical service guarantee. In particular, to achieve refined management of high-value medical consumables, many hospitals have introduced Supply-Processing-Distribution (SPD) systems to optimize the hospital's logistics supply chain. However, the application of current SPD systems in medical consumable management is still in the exploratory stage. The application of SPD systems requires demand forecasting, but existing demand forecasting methods rely heavily on the historical experience of managers and simple statistical methods (such as moving averages and year-on-year comparisons), which cannot effectively cope with the high complexity, strong uncertainty, and urgency of medical consumable demand. This leads to low forecast accuracy and a series of operational problems. On the one hand, blindly increasing inventory to avoid stockouts results in high inventory costs, serious capital occupation, and increased risk of product expiration and loss. On the other hand, due to forecast failures, emergency replenishment occurs frequently, which not only generates additional premium costs but may also affect the normal conduct of clinical diagnosis and treatment activities due to shortages of key consumables, and even endanger patient safety.

[0003] While some existing technologies have attempted to apply time series analysis, traditional machine learning models (such as BP neural networks), or single intelligent algorithms to demand forecasting, there are still some significant limitations. First, in the screening of influencing factors, there is a strong reliance on human experience (such as the Delphi method), which is highly subjective and lacks the ability to automatically and objectively identify key influencing factors from massive operational data. Second, in the construction of forecasting models, fixed-parameter models are often used, which cannot perform adaptive model optimization, have weak generalization ability, and are prone to getting trapped in local optima. In addition, in the formulation of replenishment strategies, there is a strong reliance on classic inventory control models. These models have strict assumptions and are difficult to adapt to the complex and ever-changing constraints of hospitals (such as zero stockout requirements and warehouse capacity limitations), resulting in poor performance of the formulated strategies in actual implementation.

[0004] As mentioned above, how to provide an AI-based method and system for predicting SPD consumable demand and intelligent replenishment that can automatically identify key influencing factors and formulate intelligent replenishment strategies has become an urgent problem to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide an artificial intelligence-based method and system for predicting SPD consumable demand and intelligent replenishment, in order to solve the above-mentioned problems existing in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides an artificial intelligence-based method for SPD consumable demand forecasting and intelligent replenishment, comprising: Obtain the original historical SPD consumable demand correlation data, and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. Cluster analysis is performed on the historical SPD consumable demand correlation data to generate multiple consumable demand categories. Based on each consumable demand category, SPD consumable demand influencing factor analysis is performed to identify multiple SPD consumable demand influencing factors. The correlation degree between each SPD consumable demand influencing factor and SPD consumable demand in each consumable demand category is calculated to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. An initial SPD consumable demand prediction model is obtained. Based on the key influencing factors of SPD consumables corresponding to each consumable demand category, data features are extracted from the historical SPD consumable demand correlation data to generate SPD consumable demand prediction model training data. The initial SPD consumable demand prediction model is then trained using the SPD consumable demand prediction model training data to obtain the SPD consumable demand prediction model. Obtain real-time SPD consumable demand correlation data, use the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data, obtain SPD consumable demand prediction results, and establish a replenishment objective function based on the SPD consumable demand prediction results; Based on a multi-population genetic algorithm, the replenishment objective function is iteratively optimized to obtain the optimal SPD control parameters. An intelligent SPD replenishment strategy is then generated based on the optimal SPD control parameters to control the SPD system to complete intelligent replenishment.

[0007] In one possible design, the original historical SPD consumable demand correlation data is obtained, and the historical SPD consumable demand correlation data is preprocessed to obtain historical SPD consumable demand correlation data, including: Collect raw historical SPD consumable consumption data, raw SPD consumable inventory data, raw SPD consumable attribute data, and raw SPD consumable departmental association data from the SPD system; SPD consumable entity associations are established between the original historical SPD consumable consumption data, the original SPD consumable inventory data, the original SPD consumable attribute data, and the original SPD consumable department association data to form original historical SPD consumable demand association data. The original historical SPD consumable demand association data is filled with missing values ​​and outliers are removed to obtain the pre-historical SPD consumable demand association data. Obtain SPD consumable verification rules, and perform rule verification on the pre-historical SPD consumable demand association data according to the SPD consumable verification rules to obtain pre-historical SPD consumable demand association data that passes the rule verification. The historical SPD consumable demand correlation data that has passed the rule verification is time-aligned and standardized to obtain the historical SPD consumable demand correlation data.

[0008] In one possible design, cluster analysis is performed on the historical SPD consumable demand correlation data to generate multiple consumable demand categories. Based on each consumable demand category, an analysis of SPD consumable demand influencing factors is conducted to identify multiple SPD consumable demand influencing factors. The correlation degree between each of the SPD consumable demand influencing factors and the SPD consumable demand in each consumable demand category is calculated to obtain the key SPD consumable influencing factors corresponding to each consumable demand category, including: Obtain a preset consumable category correspondence table, and perform cluster analysis on the historical SPD consumable demand association data according to the consumable category correspondence table to generate multiple consumable demand categories; Based on each of the aforementioned consumable demand categories, multiple SPD consumable demand events are generated according to the historical SPD consumable demand association data. Each SPD consumable demand event represents the association factors of an SPD consumable demand event for the current consumable demand category. Based on each SPD consumable demand event, a corresponding consumable demand category event library is established according to the consumable demand category. In the consumable demand category event library, the frequency of occurrence of the related factors in each SPD consumable demand event is calculated. Obtain a preset minimum impact frequency, and in the consumable demand category event library, filter the related factors of each SPD consumable demand event according to the minimum impact frequency, so as to obtain the related factors of SPD consumable demand events with an occurrence frequency not lower than the minimum impact frequency as the SPD consumable demand influencing factors of the current consumable demand category. The factors influencing the demand of SPD consumables for each of the consumable demand categories are extracted to form multiple sequences of SPD consumable demand influencing factors based on the event library of the consumable demand categories. Reference sequences are selected from multiple SPD consumable demand influencing factor sequences corresponding to each consumable demand category. The remaining SPD consumable demand influencing factor sequences are sorted into comparison sequences. The reference sequences and the comparison sequences are then subjected to zero-mean processing. For each of the aforementioned consumable demand categories, the correlation coefficients between the reference sequence (after zero-mean processing) and the comparison sequence (after zero-mean processing) are calculated using the following formula (1): (1) in, The first of the reference sequence The values ​​of factors influencing the demand for SPD consumables For the first The first of the comparison sequences The values ​​of factors influencing the demand for SPD consumables The preset resolution parameters, The function to find the minimum value The function for finding the maximum value. For the first The first of the comparison sequences The correlation coefficients between the factors influencing the demand for SPD consumables and the demand for SPD consumables; Based on the correlation coefficients between each SPD consumable demand influencing factor and the SPD consumable demand, the correlation degree between each of the aforementioned SPD consumable demand influencing factors and the SPD consumable demand is calculated using the following formula (2): (2) in, The total number of the comparison sequences, For the first The correlation between the factors influencing the demand for SPD consumables and the demand for SPD consumables; Obtain a preset correlation threshold. Based on the correlation threshold, in each of the consumable demand categories, select the SPD consumable demand influencing factors that are higher than the correlation threshold as the key influencing factors of SPD consumables corresponding to the consumable demand category.

[0009] In one possible design, an initial SPD consumable demand forecasting model is obtained, including: Obtain a preset SPD consumable demand prediction model structure, wherein the SPD consumable demand prediction model structure includes an input layer, a support vector regression prediction layer and an output layer connected in sequence, and the support vector regression prediction layer includes a preset kernel function. With the goal of minimizing the sum of structural risk and empirical risk, a model optimization objective function is established for the support vector regression prediction layer based on the kernel function. Based on the objective function of the model, all combinations of decision variables are calculated to form a decision variable space, wherein the combination of decision variables includes the model parameters of the support vector regression prediction layer; Obtain a preset number of population particles, and based on the number of particles, randomly define the initial particle positions and initial particle velocities in the decision variable space to form an initial particle swarm. For the decision variable space, initialize and set the individual optimal position and the global optimal position; Based on the initial particle swarm, the position and velocity of each particle are updated in the decision variable space, and the target value of the decision variable combination corresponding to the updated position of each particle is calculated. The individual optimal position and the global optimal position are updated according to the target value. When the target values ​​of each particle converge, the optimization is completed, and the final stopping position of each particle is obtained. The decision variable values ​​corresponding to the final stopping positions of each particle are used as the initial model parameters of the SPD consumable demand prediction model. The initial SPD consumable demand prediction model is set according to the initial model parameters.

[0010] In one possible design, based on the key influencing factors of SPD consumables corresponding to each of the consumable demand categories, data features are extracted from the historical SPD consumable demand correlation data to generate SPD consumable demand prediction model training data. The initial SPD consumable demand prediction model is then trained using this training data to obtain the SPD consumable demand prediction model, including: Using the key influencing factors of SPD consumables corresponding to each of the consumable demand categories as feature terms, data feature extraction is performed on the historical SPD consumable demand correlation data, and the extracted feature terms are integrated into feature vectors. Consumable demand category labels are added to each feature vector to form training data for the SPD consumable demand prediction model corresponding to each of the consumable demand categories. The training data of the SPD consumable demand prediction model is input into the input layer of the initial SPD consumable demand prediction model, and the corresponding SPD consumable demand training prediction value is output using the output layer of the initial SPD consumable demand prediction model. The parameters of the support vector regression prediction layer of the initial SPD consumable demand prediction model are adjusted using the difference between the training data of the SPD consumable demand prediction model and the training prediction value of the SPD consumable demand. Obtain the preset maximum number of training rounds, end training when the maximum number of training rounds is reached, and update the initial SPD consumable demand prediction model based on the support vector regression prediction layer obtained from the last training round to obtain the SPD consumable demand prediction model.

[0011] In one possible design, real-time SPD consumable demand correlation data is acquired, and the SPD consumable demand prediction model is used to predict the demand based on the real-time SPD consumable demand correlation data to obtain the SPD consumable demand prediction result. Based on the SPD consumable demand prediction result, a replenishment objective function is established, including: Raw real-time SPD consumable demand correlation data is collected from the SPD system, and the raw real-time SPD consumable demand correlation data is preprocessed to obtain real-time SPD consumable demand correlation data. The real-time SPD consumable demand association data is used to identify consumable demand categories. Data features are extracted from the real-time SPD consumable demand association data according to the consumable demand categories corresponding to the real-time SPD consumable demand association data. The extracted features are integrated into a real-time SPD consumable demand association data feature vector. Consumable demand category labels are added to the real-time SPD consumable demand association data feature vector to form input data. The input data is input into the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data and obtain the SPD consumable demand prediction result. Obtain the preset available funds and storage capacity boundaries. With the goal of minimizing total operating costs and minimizing stockout risk, and with the SPD consumable demand forecast results, the available funds, and the storage capacity boundaries as constraints, establish the objective optimization function for total operating costs and stockout risk through the following formula (3): (3) in, Indicates total operating costs. This indicates a risk of stockouts. The target weight coefficient, This represents the objective optimization function regarding total operating costs and stockout risk.

[0012] In one possible design, the replenishment objective function is iteratively optimized using a multi-population genetic algorithm to obtain optimal SPD control parameters. An intelligent SPD replenishment strategy is then generated based on these optimal SPD control parameters to control the SPD system for intelligent replenishment, including: Based on the replenishment objective function, all combinations of SPD control parameters are calculated, and each combination of SPD control parameters is treated as an individual to form a total individual population. The total population of individuals is divided into explorer population, developer population and elite population according to a preset ratio. The explorer population is updated by uniform mutation, the developer population is updated by Cauchy mutation, and the elite population is updated by Gaussian mutation. Obtain the preset cross-iteration rounds and the number of cross individuals. Based on the cross-iteration rounds, randomly select corresponding individuals as immigrant individuals from the explorer population, the developer population, and the elite population according to the number of cross individuals. Use the immigrant individuals to replace individuals in the explorer population, the developer population, and the elite population. The objective function value of each individual in the explorer population, the developer population, and the elite population after each iteration update is calculated respectively. The objective function value of each individual in the explorer population, the developer population, and the elite population is calculated respectively. Obtain a preset genetic iteration round, perform multiple rounds of iterative updates and cross-iterations on the explorer population, the developer population, and the elite population, and end the iterative update after reaching the genetic iteration round. Calculate the objective function value for each individual in the elite population after the last iteration update, and select the individual with the highest objective function value as the optimal individual. The optimal SPD control parameters are the combination of SPD control parameters corresponding to the optimal individual. Obtain the current SPD control parameters, and generate an SPD control parameter adjustment strategy based on the difference between the optimal SPD control parameters and the SPD control parameters. Use the SPD control parameter adjustment strategy as an intelligent SPD replenishment strategy, and control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

[0013] Secondly, the present invention provides an artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment system, comprising: The SPD consumable data acquisition unit is used to acquire the original historical SPD consumable demand correlation data and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. The influencing factor analysis unit is used to perform cluster analysis on the historical SPD consumable demand correlation data, generate multiple consumable demand categories, and perform SPD consumable demand influencing factor analysis based on each consumable demand category to identify multiple SPD consumable demand influencing factors. The correlation degree of each SPD consumable demand influencing factor in each consumable demand category to SPD consumable demand is calculated to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. The demand forecasting model generation unit is used to obtain an initial SPD consumable demand forecasting model, extract data features from the historical SPD consumable demand correlation data according to the key influencing factors of SPD consumables corresponding to each consumable demand category, generate SPD consumable demand forecasting model training data, and use the SPD consumable demand forecasting model training data to train the initial SPD consumable demand forecasting model to obtain the SPD consumable demand forecasting model. The SPD consumable demand real-time forecasting unit is used to acquire real-time SPD consumable demand related data, use the SPD consumable demand forecasting model to forecast the demand of the real-time SPD consumable demand related data, obtain the SPD consumable demand forecasting result, and establish a replenishment objective function based on the SPD consumable demand forecasting result. The SPD replenishment strategy generation unit is used to iteratively optimize the replenishment objective function based on a multi-population genetic algorithm to obtain the optimal SPD control parameters, and generate an intelligent SPD replenishment strategy based on the optimal SPD control parameters, so as to control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

[0014] Thirdly, the present invention provides an electronic device comprising a memory, a processor, and a transceiver connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method as described in the first aspect or any possible design of the first aspect.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform the AI-based SPD consumable demand forecasting and intelligent replenishment method described in the first aspect or any possible design of the first aspect.

[0016] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the AI-based SPD consumable demand forecasting and intelligent replenishment method as described in the first aspect or any possible design of the first aspect.

[0017] Beneficial Effects: This invention provides an artificial intelligence-based method and system for SPD consumable demand forecasting and intelligent replenishment, comprising: First, acquiring original historical SPD consumable demand correlation data and performing data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data; Second, performing cluster analysis on the historical SPD consumable demand correlation data to generate multiple consumable demand categories, and performing SPD consumable demand influencing factor analysis based on each consumable demand category to identify multiple SPD consumable demand influencing factors, calculating the correlation degree of each SPD consumable demand influencing factor in each consumable demand category with the SPD consumable demand, so as to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category; Then, acquiring an initial SPD consumable demand forecasting model, and based on the SPD consumable demand corresponding to each consumable demand category... The key influencing factors of SPD consumables are analyzed. Data features are extracted from historical SPD consumable demand correlation data to generate training data for an SPD consumable demand prediction model. This training data is then used to train the initial SPD consumable demand prediction model, resulting in an SPD consumable demand prediction model. Next, real-time SPD consumable demand correlation data is acquired, and the SPD consumable demand prediction model is used to predict demand based on this data, yielding SPD consumable demand prediction results. A replenishment objective function is then established based on these results. Finally, using a multi-population genetic algorithm, the replenishment objective function is iteratively optimized to obtain optimal SPD control parameters. An intelligent SPD replenishment strategy is generated based on these optimal parameters to control the SPD system for intelligent replenishment. By clustering analysis of historical SPD consumable demand correlation data and identifying influencing factors of SPD consumable demand, correlation degree calculation is performed to obtain key influencing factors of SPD consumables. Furthermore, by acquiring and training an initial SPD consumable demand prediction model, an accurate SPD consumable demand prediction model is obtained. This model is then used to predict real-time SPD consumable demand correlation data, outputting SPD consumable demand prediction results that better match actual needs and applications. In addition, the replenishment objective function is iteratively optimized using a multi-population genetic algorithm to obtain the optimal SPD control parameters. These optimal SPD control parameters enable the automated generation of intelligent replenishment strategies. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the AI-based SPD consumable demand forecasting and intelligent replenishment method provided in this embodiment of the invention. Figure 2 A schematic diagram of the functional structure of the AI-based SPD consumable demand forecasting and intelligent replenishment system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0020] It should be understood that although the terms first, second, etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, without departing from the scope of the exemplary embodiments of the invention.

[0021] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0022] Example: like Figure 1 As shown, the first aspect of this embodiment provides an artificial intelligence-based method for SPD consumable demand forecasting and intelligent replenishment, which may include, but is not limited to, the following steps: S1. Obtain the original historical SPD consumable demand association data, and perform data preprocessing on the historical SPD consumable demand association data to obtain historical SPD consumable demand association data; S2. Perform cluster analysis on the historical SPD consumable demand correlation data to generate multiple consumable demand categories, and perform SPD consumable demand influencing factor analysis based on each consumable demand category to identify multiple SPD consumable demand influencing factors. Calculate the correlation degree between each SPD consumable demand influencing factor and SPD consumable demand in each consumable demand category to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. S3. Obtain an initial SPD consumable demand prediction model. Based on the key influencing factors of SPD consumables corresponding to each consumable demand category, extract data features from the historical SPD consumable demand correlation data to generate SPD consumable demand prediction model training data. Use the SPD consumable demand prediction model training data to train the initial SPD consumable demand prediction model to obtain the SPD consumable demand prediction model. S4. Obtain real-time SPD consumable demand correlation data, use the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data, obtain SPD consumable demand prediction results, and establish a replenishment objective function based on the SPD consumable demand prediction results. S5. Based on a multi-population genetic algorithm, the replenishment objective function is iteratively optimized to obtain the optimal SPD control parameters. An intelligent SPD replenishment strategy is generated based on the optimal SPD control parameters, so as to control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

[0023] In one possible implementation, step S1 involves acquiring the original historical SPD consumable demand correlation data and performing data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. This can be broken down into, but is not limited to, the following steps S11-S15, specifically including: S11. Collect original historical SPD consumable consumption data, original SPD consumable inventory data, original SPD consumable attribute data, and original SPD consumable departmental association data from the SPD system; S12. Perform SPD consumable entity association between the original historical SPD consumable consumption data, the original SPD consumable inventory data, the original SPD consumable attribute data, and the original SPD consumable department association data to form original historical SPD consumable demand association data; S13. Fill missing values ​​and remove outliers from the original historical SPD consumable demand association data to obtain pre-historical SPD consumable demand association data; S14. Obtain SPD consumable verification rules, and perform rule verification on the pre-historical SPD consumable demand association data according to the SPD consumable verification rules to obtain pre-historical SPD consumable demand association data that has passed the rule verification. S15. Perform time alignment and standardization on the pre-historical SPD consumable demand association data that has passed the rule verification to obtain historical SPD consumable demand association data.

[0024] It should be noted that the AI-based SPD consumable demand forecasting and intelligent replenishment method provided in this embodiment establishes an automated collection and standardized preprocessing pipeline for multi-source, heterogeneous SPD data, laying a solid data foundation for subsequent intelligent analysis. Through direct interface with the SPD system and data collection, it integrates multi-dimensional SPD-related data, including but not limited to SPD consumable consumption data, SPD consumable inventory data, SPD consumable attribute data, SPD consumable departmental related data, and external environment related data, forming complete SPD consumable demand-related data. Furthermore, through automated time alignment and standardized processing (including but not limited to data cleaning and anomaly detection), the quality of the original SPD consumable demand-related data is improved from experience-dependent to a quantifiable and controllable standard level. The average data availability rate increases from approximately 70% with manual processing to over 98%, significantly reducing the negative impact of abnormal data on subsequent demand forecasting models.

[0025] In one possible implementation, step S2 involves performing cluster analysis on the historical SPD consumable demand correlation data to generate multiple consumable demand categories. Based on each consumable demand category, an analysis of SPD consumable demand influencing factors is conducted to identify multiple SPD consumable demand influencing factors. The correlation degree between each SPD consumable demand influencing factor and the SPD consumable demand in each consumable demand category is calculated to obtain the key influencing factors for SPD consumables corresponding to each consumable demand category. This can be decomposed, but is not limited to, the following steps S21-S29, specifically including: S21. Obtain a preset consumable category correspondence table, and perform cluster analysis on the historical SPD consumable demand association data according to the consumable category correspondence table to generate multiple consumable demand categories; S22. Based on each of the consumable demand categories, generate multiple SPD consumable demand events according to the historical SPD consumable demand association data, wherein the SPD consumable demand event represents the association factors of an SPD consumable demand event for the current consumable demand category; S23. Based on each SPD consumable demand event, establish a corresponding consumable demand category event library according to the consumable demand category, and calculate the frequency of occurrence of the related factors in each SPD consumable demand event in the consumable demand category event library; S24. Obtain a preset minimum impact frequency. In the consumable demand category event library, filter the related factors of each SPD consumable demand event according to the minimum impact frequency, so as to obtain the related factors of SPD consumable demand events with an occurrence frequency not lower than the minimum impact frequency as the SPD consumable demand influencing factors of the current consumable demand category. S25. Extract the SPD consumable demand influencing factors for each of the consumable demand categories, and form multiple corresponding SPD consumable demand influencing factor sequences based on the consumable demand category event library; S26. Select a reference sequence from the multiple SPD consumable demand influencing factor sequences corresponding to each consumable demand category, sort the remaining SPD consumable demand influencing factor sequences into comparison sequences, and perform zero-mean processing on the reference sequence and the comparison sequence. S27. For each of the aforementioned consumable demand categories, the correlation coefficient between the reference sequence (after zero-mean processing) and the comparison sequence (after zero-mean processing) is calculated using the following formula (1): (1) in, The first of the reference sequence The values ​​of factors influencing the demand for SPD consumables For the first The first of the comparison sequences The values ​​of factors influencing the demand for SPD consumables The preset resolution parameters, The function to find the minimum value The function for finding the maximum value. For the first The first of the comparison sequences The correlation coefficients between the factors influencing the demand for SPD consumables and the demand for SPD consumables; S28. Based on the correlation coefficients of each SPD consumable demand influencing factor to the SPD consumable demand, the correlation degree of each of the SPD consumable demand influencing factors to the SPD consumable demand is calculated using the following formula (2): (2) in, The total number of the comparison sequences, For the first The correlation between the factors influencing the demand for SPD consumables and the demand for SPD consumables; S29. Obtain a preset correlation threshold. Based on the correlation threshold, in each of the consumable demand categories, select the SPD consumable demand influencing factors that are higher than the correlation threshold as the key influencing factors of SPD consumables corresponding to the consumable demand category.

[0026] It should be noted that the AI-based SPD consumable demand forecasting and intelligent replenishment method provided in this embodiment completely changes the traditional qualitative analysis method that relies on expert experience through a data-driven intelligent screening mechanism. It achieves the objectification and quantitative identification of SPD consumable demand influencing factors, and calculates the correlation between SPD consumable demand influencing factors and SPD consumable demand. It automatically discovers key SPD consumable influencing factors related to SPD consumable demand from massive historical SPD consumable demand correlation data, avoiding the subjective bias and omission of important factors in manual screening. In simulation experiments, the accuracy of the screened key SPD consumable influencing factors is 40% higher than that of the traditional Delphi method. Furthermore, based on the analysis and identification of consumable demand categories, a differentiated influencing factor system is formed. For example, it is identified that surgical consumables are mainly affected by the number of surgeries and the development of specialties, while ordinary consumables are more affected by outpatient volume and seasonal factors. This refined management of influencing factors enables the subsequent demand forecasting model to have complete predictive accuracy for different consumable demand categories.

[0027] In one possible implementation, step S3, obtaining the initial SPD consumable demand prediction model, can be decomposed into, but is not limited to, the following steps S3a1-S3a8, specifically including: S3a1. Obtain a preset SPD consumable demand prediction model structure, wherein the SPD consumable demand prediction model structure includes an input layer, a support vector regression prediction layer and an output layer connected in sequence, and the support vector regression prediction layer includes a preset kernel function. S3a2. With the goal of minimizing the sum of structural risk and empirical risk, a model optimization objective function is established for the support vector regression prediction layer based on the kernel function; S3a3. Optimize the objective function according to the model, calculate all combinations of decision variables, and form a decision variable space based on all combinations of decision variables, wherein the combination of decision variables includes the model parameters of the support vector regression prediction layer; S3a4. Obtain the preset number of population particles, and according to the number of particles, randomly define the initial particle position and initial particle velocity in the decision variable space to form an initial particle swarm; S3a5. Initialize and set the individual optimal position and the global optimal position for the decision variable space; S3a6. Based on the initial particle swarm, update the position and velocity of each particle in the decision variable space, calculate the target value of the decision variable combination corresponding to the position of each particle after the update, and update the individual optimal position and the global optimal position according to the target value; S3a7. When the target values ​​of each particle converge, the optimization is completed, and the final stopping position of each particle is obtained. The decision variable values ​​corresponding to the final stopping positions of each particle are used as the initial model parameters of the SPD consumable demand prediction model. S3a8. Set the initial SPD consumable demand prediction model according to the initial model parameters.

[0028] In one possible implementation, in step S3, based on the key influencing factors of SPD consumables corresponding to each of the consumable demand categories, data features are extracted from the historical SPD consumable demand correlation data to generate SPD consumable demand prediction model training data. The initial SPD consumable demand prediction model is then trained using this training data to obtain the SPD consumable demand prediction model. This can be decomposed, but is not limited to, the following steps S3b1-S3b4, specifically including: S3b1. Using the key influencing factors of SPD consumables corresponding to each of the consumable demand categories as feature terms, extract data features from the historical SPD consumable demand correlation data, integrate the extracted feature terms into feature vectors, add consumable demand category labels to each feature vector, so as to form training data for the SPD consumable demand prediction model corresponding to each of the consumable demand categories. S3b2. Input the training data of the SPD consumable demand prediction model into the input layer of the initial SPD consumable demand prediction model, and output the corresponding SPD consumable demand training prediction value using the output layer of the initial SPD consumable demand prediction model. S3b3. Adjust the parameters of the support vector regression prediction layer of the initial SPD consumable demand prediction model using the difference between the training data of the SPD consumable demand prediction model and the training prediction value of the SPD consumable demand. S3b4. Obtain the preset maximum number of training rounds, end training when the maximum number of training rounds is reached, and update the initial SPD consumable demand prediction model based on the support vector regression prediction layer obtained from the last training round to obtain the SPD consumable demand prediction model.

[0029] The AI-based SPD consumable demand forecasting and intelligent replenishment method provided in this embodiment deeply integrates Particle Swarm Optimization (PSO) and Support Vector Regression (SVR), achieving a breakthrough improvement in prediction accuracy and efficiency. In practical applications, the PSO algorithm can be further optimized using Improved Particle Swarm Optimization (IPSO). Through adaptive inertia weights and dynamic learning factors, the parameter optimization process achieves an optimal balance between exploration and development. Compared with traditional grid search and standard PSO algorithms, the optimization speed is significantly improved, and it can effectively avoid getting trapped in local optima. Ultimately, the root mean square error of the SVR algorithm is reduced, thereby realizing dynamic parameter optimization of the initial SPD consumable demand forecasting model to obtain a more accurate and realistic SPD consumable demand forecasting model, significantly improving the model performance and prediction accuracy of the SPD consumable demand forecasting model.

[0030] In one possible implementation, step S4 involves acquiring real-time SPD consumable demand correlation data, using the SPD consumable demand prediction model to predict the demand based on the real-time SPD consumable demand correlation data, obtaining the SPD consumable demand prediction result, and establishing a replenishment objective function based on the SPD consumable demand prediction result. This can be decomposed into, but is not limited to, the following steps S41-S44, specifically including: S41. Collect raw real-time SPD consumable demand correlation data from the SPD system, and perform data preprocessing on the raw real-time SPD consumable demand correlation data to obtain real-time SPD consumable demand correlation data. S42. Perform consumable demand category identification on the real-time SPD consumable demand association data, so as to extract data features from the real-time SPD consumable demand association data according to the consumable demand category corresponding to the real-time SPD consumable demand association data, integrate the extracted feature items into a real-time SPD consumable demand association data feature vector, and add consumable demand category labels to the real-time SPD consumable demand association data feature vector to form input data; S43. Input the input data into the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data and obtain the SPD consumable demand prediction result. S44. Obtain the preset available funds and storage capacity boundary. With the goal of minimizing total operating costs and minimizing stockout risk, and with the SPD consumable demand forecast, available funds, and storage capacity boundary as constraints, establish the objective optimization function for total operating costs and stockout risk using the following formula (3): (3) in, Indicates total operating costs. This indicates a risk of stockouts. The target weight coefficient, This represents the objective optimization function regarding total operating costs and stockout risk.

[0031] In one possible implementation, step S5 involves iteratively optimizing the replenishment objective function based on a multi-population genetic algorithm to obtain optimal SPD control parameters. An intelligent SPD replenishment strategy is then generated based on these optimal SPD control parameters to control the SPD system for intelligent replenishment. This step can be decomposed, but is not limited to, the following steps S51-S57, specifically including: S51. Based on the replenishment objective function, calculate all combinations of SPD control parameters, and treat each combination of SPD control parameters as an individual to form a total individual population; S52. Divide the total individual population into an explorer population, a developer population, and an elite population according to a preset ratio. Perform uniform mutation update on the explorer population, Cauchy mutation update on the developer population, and Gaussian mutation update on the elite population. S53. Obtain the preset cross-iteration rounds and the number of cross individuals. Based on the cross-iteration rounds, randomly select corresponding individuals as immigrant individuals from the explorer population, the developer population, and the elite population according to the number of cross individuals. Use the immigrant individuals to replace individuals in the explorer population, the developer population, and the elite population. S54. Calculate the objective function value of each individual in the explorer population, the developer population, and the elite population after each iteration update, and calculate the objective function value of each individual in the explorer population, the developer population, and the elite population respectively; S55. Obtain a preset genetic iteration round, perform multiple rounds of iterative updates and cross-iterations on the explorer population, the developer population, and the elite population, and end the iterative update after reaching the genetic iteration round. Calculate the objective function value for each individual in the elite population after the last iteration update, and select the individual with the highest objective function value as the optimal individual. S56. The combination of SPD control parameters corresponding to the optimal individual is taken as the optimal SPD control parameters; S57. Obtain the current SPD control parameters, generate an SPD control parameter adjustment strategy based on the difference between the optimal SPD control parameters and the SPD control parameters, use the SPD control parameter adjustment strategy as an intelligent SPD replenishment strategy, and control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

[0032] It should be noted that the AI-based SPD consumable demand forecasting and intelligent replenishment method provided in this embodiment solves a multi-objective optimization problem by using a multi-population genetic algorithm to address the constraints formed by the SPD consumable demand forecasting results. This achieves the scientific and precise formulation of replenishment strategies. First, the multi-population parallel architecture of the multi-population genetic algorithm can simultaneously perform deep searches in different regions of the strategy space, improving the optimization efficiency of finding the global optimum compared to a single-population genetic algorithm. Furthermore, the final generated strategy can reduce the total operating cost by an average of 18% to 25% in simulation tests. In addition, the multi-population genetic algorithm fully considers the actual constraints in hospital operations (such as the available funds and the warehouse capacity boundary), and the generated strategy can fully meet the feasibility of the business and maintain the expected performance under various demand scenarios. Moreover, its optimization results not only provide static optimal SPD control parameters but also output condition-triggered dynamic intelligent SPD replenishment strategies. It can also automatically adjust the safety stock coefficient according to the magnitude of the prediction error, enabling the SPD system to have adaptive adjustment capabilities.

[0033] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method described in the first aspect of the embodiment, including: The SPD consumable data acquisition unit is used to acquire the original historical SPD consumable demand correlation data and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. The influencing factor analysis unit is used to perform cluster analysis on the historical SPD consumable demand correlation data, generate multiple consumable demand categories, and perform SPD consumable demand influencing factor analysis based on each consumable demand category to identify multiple SPD consumable demand influencing factors. The correlation degree of each SPD consumable demand influencing factor in each consumable demand category to SPD consumable demand is calculated to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. The demand forecasting model generation unit is used to obtain an initial SPD consumable demand forecasting model, extract data features from the historical SPD consumable demand correlation data according to the key influencing factors of SPD consumables corresponding to each consumable demand category, generate SPD consumable demand forecasting model training data, and use the SPD consumable demand forecasting model training data to train the initial SPD consumable demand forecasting model to obtain the SPD consumable demand forecasting model. The SPD consumable demand real-time forecasting unit is used to acquire real-time SPD consumable demand related data, use the SPD consumable demand forecasting model to forecast the demand of the real-time SPD consumable demand related data, obtain the SPD consumable demand forecasting result, and establish a replenishment objective function based on the SPD consumable demand forecasting result. The SPD replenishment strategy generation unit is used to iteratively optimize the replenishment objective function based on a multi-population genetic algorithm to obtain the optimal SPD control parameters, and generate an intelligent SPD replenishment strategy based on the optimal SPD control parameters, so as to control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

[0034] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0035] like Figure 3 As shown, the third aspect of this embodiment provides an electronic device, including: a memory, a processor, and a transceiver that are sequentially and communicatively connected, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method as described in the first aspect of the embodiment.

[0036] For specific examples, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; specifically, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), PLA (Programmable Logic Array). The processor may also include a main processor and a coprocessor. The main processor, also known as the CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state.

[0037] In some embodiments, the processor may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. For example, the processor may not be limited to microprocessors of the STM32F105 series, reduced instruction set computer (RISC) microprocessors, x86 architecture processors, or processors with integrated neural network processing units (NPUs). The transceiver may be, but is not limited to, a Wi-Fi transceiver, a Bluetooth transceiver, a General Packet Radio Service (GPRS) transceiver, a ZigBee transceiver (a low-power LAN protocol based on the IEEE 802.15.4 standard), a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. Furthermore, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0038] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0039] The fourth aspect of this embodiment provides a storage medium that stores instructions containing the AI-based SPD consumable demand forecasting and intelligent replenishment method described in the first aspect of the embodiment. That is, the storage medium stores instructions that, when executed on a computer, perform the AI-based SPD consumable demand forecasting and intelligent replenishment method as described in the first aspect of the embodiment.

[0040] The storage medium refers to a carrier for storing data, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0041] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0042] The fifth aspect of this embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the AI-based SPD consumable demand forecasting and intelligent replenishment method as described in the first aspect of this embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0043] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for demand forecasting and intelligent replenishment of SPD consumables based on artificial intelligence, characterized in that, include: Obtain the original historical SPD consumable demand correlation data, and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. Cluster analysis is performed on the historical SPD consumable demand correlation data to generate multiple consumable demand categories. Based on each consumable demand category, SPD consumable demand influencing factor analysis is performed to identify multiple SPD consumable demand influencing factors. The correlation degree between each SPD consumable demand influencing factor and SPD consumable demand in each consumable demand category is calculated to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. An initial SPD consumable demand prediction model is obtained. Based on the key influencing factors of SPD consumables corresponding to each consumable demand category, data features are extracted from the historical SPD consumable demand correlation data to generate SPD consumable demand prediction model training data. The initial SPD consumable demand prediction model is then trained using the SPD consumable demand prediction model training data to obtain the SPD consumable demand prediction model. Obtain real-time SPD consumable demand correlation data, use the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data, obtain SPD consumable demand prediction results, and establish a replenishment objective function based on the SPD consumable demand prediction results; Based on a multi-population genetic algorithm, the replenishment objective function is iteratively optimized to obtain the optimal SPD control parameters. An intelligent SPD replenishment strategy is then generated based on the optimal SPD control parameters to control the SPD system to complete intelligent replenishment.

2. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 1, characterized in that, Obtain the original historical SPD consumable demand correlation data, and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data, including: Collect raw historical SPD consumable consumption data, raw SPD consumable inventory data, raw SPD consumable attribute data, and raw SPD consumable departmental association data from the SPD system; SPD consumable entity associations are established between the original historical SPD consumable consumption data, the original SPD consumable inventory data, the original SPD consumable attribute data, and the original SPD consumable department association data to form original historical SPD consumable demand association data. The original historical SPD consumable demand association data is filled with missing values ​​and outliers are removed to obtain the pre-historical SPD consumable demand association data. Obtain SPD consumable verification rules, and perform rule verification on the pre-historical SPD consumable demand association data according to the SPD consumable verification rules to obtain pre-historical SPD consumable demand association data that passes the rule verification. The historical SPD consumable demand correlation data that has passed the rule verification is time-aligned and standardized to obtain the historical SPD consumable demand correlation data.

3. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 1, characterized in that, Cluster analysis is performed on the historical SPD consumable demand correlation data to generate multiple consumable demand categories. Based on each consumable demand category, an analysis of SPD consumable demand influencing factors is conducted to identify multiple SPD consumable demand influencing factors. The correlation degree between each of the SPD consumable demand influencing factors and SPD consumable demand in each consumable demand category is calculated to obtain the key SPD consumable influencing factors corresponding to each consumable demand category, including: Obtain a preset consumable category correspondence table, and perform cluster analysis on the historical SPD consumable demand association data according to the consumable category correspondence table to generate multiple consumable demand categories; Based on each of the aforementioned consumable demand categories, multiple SPD consumable demand events are generated according to the historical SPD consumable demand association data. Each SPD consumable demand event represents the association factors of an SPD consumable demand event for the current consumable demand category. Based on each SPD consumable demand event, a corresponding consumable demand category event library is established according to the consumable demand category. In the consumable demand category event library, the frequency of occurrence of the related factors in each SPD consumable demand event is calculated. Obtain a preset minimum impact frequency, and in the consumable demand category event library, filter the related factors of each SPD consumable demand event according to the minimum impact frequency, so as to obtain the related factors of SPD consumable demand events with an occurrence frequency not lower than the minimum impact frequency as the SPD consumable demand influencing factors of the current consumable demand category. The factors influencing the demand of SPD consumables for each of the consumable demand categories are extracted to form multiple sequences of SPD consumable demand influencing factors based on the event library of the consumable demand categories. Reference sequences are selected from multiple SPD consumable demand influencing factor sequences corresponding to each consumable demand category. The remaining SPD consumable demand influencing factor sequences are sorted into comparison sequences. The reference sequences and the comparison sequences are then subjected to zero-mean processing. For each of the aforementioned consumable demand categories, the correlation coefficients between the reference sequence (after zero-mean processing) and the comparison sequence (after zero-mean processing) are calculated using the following formula (1): (1) in, The first of the reference sequence The values ​​of factors influencing the demand for SPD consumables For the first The first of the comparison sequences The values ​​of factors influencing the demand for SPD consumables The preset resolution parameters, The function to find the minimum value The function for finding the maximum value. For the first The first of the comparison sequences The correlation coefficients of the factors influencing the demand for SPD consumables on the demand for SPD consumables; Based on the correlation coefficients between each SPD consumable demand influencing factor and the SPD consumable demand, the correlation degree between each of the aforementioned SPD consumable demand influencing factors and the SPD consumable demand is calculated using the following formula (2): (2) in, The total number of the comparison sequences, For the first The correlation between the factors influencing the demand for SPD consumables and the demand for SPD consumables; Obtain a preset correlation threshold. Based on the correlation threshold, in each of the consumable demand categories, select the SPD consumable demand influencing factors that are higher than the correlation threshold as the key influencing factors of SPD consumables corresponding to the consumable demand category.

4. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 1, characterized in that, Obtain the initial SPD consumable demand forecasting model, including: Obtain a preset SPD consumable demand prediction model structure, wherein the SPD consumable demand prediction model structure includes an input layer, a support vector regression prediction layer and an output layer connected in sequence, and the support vector regression prediction layer includes a preset kernel function. With the goal of minimizing the sum of structural risk and empirical risk, a model optimization objective function is established for the support vector regression prediction layer based on the kernel function. Based on the objective function of the model, all combinations of decision variables are calculated to form a decision variable space, wherein the combination of decision variables includes the model parameters of the support vector regression prediction layer; Obtain a preset number of population particles, and based on the number of particles, randomly define the initial particle positions and initial particle velocities in the decision variable space to form an initial particle swarm. For the decision variable space, initialize and set the individual optimal position and the global optimal position; Based on the initial particle swarm, the position and velocity of each particle are updated in the decision variable space, and the target value of the decision variable combination corresponding to the updated position of each particle is calculated. The individual optimal position and the global optimal position are updated according to the target value. When the target values ​​of each particle converge, the optimization is completed, and the final stopping position of each particle is obtained. The decision variable values ​​corresponding to the final stopping positions of each particle are used as the initial model parameters of the SPD consumable demand prediction model. The initial SPD consumable demand prediction model is set according to the initial model parameters.

5. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 4, characterized in that, Based on the key influencing factors of SPD consumables corresponding to each of the consumable demand categories, data features are extracted from the historical SPD consumable demand correlation data to generate training data for the SPD consumable demand prediction model. The initial SPD consumable demand prediction model is then trained using this training data to obtain the SPD consumable demand prediction model, which includes: Using the key influencing factors of SPD consumables corresponding to each of the consumable demand categories as feature terms, data feature extraction is performed on the historical SPD consumable demand correlation data, and the extracted feature terms are integrated into feature vectors. Consumable demand category labels are added to each feature vector to form training data for the SPD consumable demand prediction model corresponding to each of the consumable demand categories. The training data of the SPD consumable demand prediction model is input into the input layer of the initial SPD consumable demand prediction model, and the corresponding SPD consumable demand training prediction value is output using the output layer of the initial SPD consumable demand prediction model. The parameters of the support vector regression prediction layer of the initial SPD consumable demand prediction model are adjusted using the difference between the training data of the SPD consumable demand prediction model and the training prediction value of the SPD consumable demand. Obtain the preset maximum number of training rounds, end training when the maximum number of training rounds is reached, and update the initial SPD consumable demand prediction model based on the support vector regression prediction layer obtained from the last training round to obtain the SPD consumable demand prediction model.

6. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 1, characterized in that, Obtain real-time SPD consumable demand correlation data, use the SPD consumable demand prediction model to predict the demand based on the real-time SPD consumable demand correlation data, obtain the SPD consumable demand prediction result, and establish a replenishment objective function based on the SPD consumable demand prediction result, including: Raw real-time SPD consumable demand correlation data is collected from the SPD system, and the raw real-time SPD consumable demand correlation data is preprocessed to obtain real-time SPD consumable demand correlation data. The real-time SPD consumable demand association data is used to identify consumable demand categories. Data features are extracted from the real-time SPD consumable demand association data according to the consumable demand categories corresponding to the real-time SPD consumable demand association data. The extracted features are integrated into a real-time SPD consumable demand association data feature vector. Consumable demand category labels are added to the real-time SPD consumable demand association data feature vector to form input data. The input data is input into the SPD consumable demand prediction model to predict the demand of the real-time SPD consumable demand correlation data and obtain the SPD consumable demand prediction result. Obtain the preset available funds and storage capacity boundaries. With the goal of minimizing total operating costs and minimizing stockout risk, and with the SPD consumable demand forecast results, the available funds, and the storage capacity boundaries as constraints, establish the objective optimization function for total operating costs and stockout risk through the following formula (3): (3) in, Indicates total operating costs. This indicates a risk of stockouts. The target weight coefficient, This represents the objective optimization function regarding total operating costs and stockout risk.

7. The artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method according to claim 1, characterized in that, Based on a multi-population genetic algorithm, the replenishment objective function is iteratively optimized to obtain the optimal SPD control parameters. An intelligent SPD replenishment strategy is then generated based on these optimal SPD control parameters to control the SPD system for intelligent replenishment, including: Based on the replenishment objective function, all combinations of SPD control parameters are calculated, and each combination of SPD control parameters is treated as an individual to form a total individual population. The total population of individuals is divided into explorer population, developer population and elite population according to a preset ratio. The explorer population is updated by uniform mutation, the developer population is updated by Cauchy mutation, and the elite population is updated by Gaussian mutation. Obtain the preset cross-iteration rounds and the number of cross individuals. Based on the cross-iteration rounds, randomly select corresponding individuals as immigrant individuals from the explorer population, the developer population, and the elite population according to the number of cross individuals. Use the immigrant individuals to replace individuals in the explorer population, the developer population, and the elite population. The objective function value of each individual in the explorer population, the developer population, and the elite population after each iteration update is calculated respectively. The objective function value of each individual in the explorer population, the developer population, and the elite population is calculated respectively. Obtain a preset genetic iteration round, perform multiple rounds of iterative updates and cross-iterations on the explorer population, the developer population, and the elite population, and end the iterative update after reaching the genetic iteration round. Calculate the objective function value for each individual in the elite population after the last iteration update, and select the individual with the highest objective function value as the optimal individual. The optimal SPD control parameters are the combination of SPD control parameters corresponding to the optimal individual. Obtain the current SPD control parameters, and generate an SPD control parameter adjustment strategy based on the difference between the optimal SPD control parameters and the SPD control parameters. Use the SPD control parameter adjustment strategy as an intelligent SPD replenishment strategy, and control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

8. An AI-based SPD consumable demand forecasting and intelligent replenishment system, characterized in that, The method for predicting and intelligently replenishing SPD consumable demand based on artificial intelligence as described in any one of claims 1 to 7 includes: The SPD consumable data acquisition unit is used to acquire the original historical SPD consumable demand correlation data and perform data preprocessing on the historical SPD consumable demand correlation data to obtain historical SPD consumable demand correlation data. The influencing factor analysis unit is used to perform cluster analysis on the historical SPD consumable demand correlation data, generate multiple consumable demand categories, and perform SPD consumable demand influencing factor analysis based on each consumable demand category to identify multiple SPD consumable demand influencing factors. The correlation degree of each SPD consumable demand influencing factor in each consumable demand category to SPD consumable demand is calculated to obtain the key influencing factors of SPD consumables corresponding to each consumable demand category. The demand forecasting model generation unit is used to obtain an initial SPD consumable demand forecasting model, extract data features from the historical SPD consumable demand correlation data according to the key influencing factors of SPD consumables corresponding to each consumable demand category, generate SPD consumable demand forecasting model training data, and use the SPD consumable demand forecasting model training data to train the initial SPD consumable demand forecasting model to obtain the SPD consumable demand forecasting model. The SPD consumable demand real-time forecasting unit is used to acquire real-time SPD consumable demand related data, use the SPD consumable demand forecasting model to forecast the demand of the real-time SPD consumable demand related data, obtain the SPD consumable demand forecasting result, and establish a replenishment objective function based on the SPD consumable demand forecasting result. The SPD replenishment strategy generation unit is used to iteratively optimize the replenishment objective function based on a multi-population genetic algorithm to obtain the optimal SPD control parameters, and generate an intelligent SPD replenishment strategy based on the optimal SPD control parameters, so as to control the SPD system to complete intelligent replenishment according to the intelligent SPD replenishment strategy.

9. An electronic device, characterized in that, The device includes a memory, a processor, and a transceiver that are sequentially and communicatively connected. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the AI-based SPD consumable demand forecasting and intelligent replenishment method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the artificial intelligence-based SPD consumable demand forecasting and intelligent replenishment method as described in any one of claims 1 to 7.