Production line capacity intelligent management method for pharmaceutical production

Abstract

The invention relates to the technical field of intelligent productivity management, and particularly discloses an intelligent production line productivity management method for pharmaceutical production, which comprises the following steps of: S1, setting a historical data acquisition interval, and obtaining sales volume and corresponding morbidity of all indications of a drug; establishing a basic relational expression between the sales volume and the morbidity through a regression model; s2, after data of abnormal years are removed, incidence curves are generated according to years, and periodic division is carried out; for each morbidity cycle, the same-period curves over the years are aggregated to form a set, year difference and curve similarity are analyzed through pairwise combination, year weight coefficients are calculated, and finally a standard morbidity curve is generated through weighted fusion; s3, monitoring and generating a current morbidity curve in real time in a new morbidity period, and comparing the current morbidity curve with a standard curve to predict the morbidity of a next node; and in combination with the basic relational expression, converting the predicted morbidity into recommended yield, and realizing dynamic productivity planning.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent capacity management, and in particular to a method for intelligent capacity management of a production line for drug production. Background Art

[0002] Pharmaceutical capacity management refers to the management and control of the production quantity and quality of drugs by pharmaceutical companies during the production process; it mainly includes production planning and scheduling, equipment management, raw material management, quality management and supply chain management.

[0003] In the existing drug production technology system, drug production is usually arranged according to pre-set production standards. Although this relatively fixed and standardized production method can ensure the planning and stability of production to a certain extent, it has significant limitations, especially in the face of dynamically changing market environments and patient needs, its drawbacks become more and more prominent.

[0004] Specifically, this traditional production method lacks sufficient flexibility and is difficult to make timely and adaptive adjustments based on the current complex and changing environment and the actual incidence of the indications. In different regions, seasons, and even years, the incidence of various diseases may fluctuate due to a variety of factors. For example, for some diseases with seasonal onset characteristics, the number of patients may increase sharply during the high-incidence season, and the demand for related drugs will also increase significantly accordingly; in some special cases, such as public health emergencies or localized disease epidemics, the demand for specific drugs may explode. However, the production model based on preset production standards cannot keenly capture these changes and cannot quickly optimize production.

[0005] For drugs with a short shelf life, this situation poses even more serious problems. On the one hand, if the market demand is not accurately estimated during the production process, resulting in overproduction, a large number of drugs will face the fate of expiration if they cannot be fully sold and used within the validity period of the drugs. This not only means that the raw materials, manpower, material resources and other resources invested in the production process are wasted for no reason, but may also have a negative impact on the economic benefits of the enterprise. On the other hand, if the output is underestimated and lower than the actual market demand, then when patients urgently need the drug for treatment, there may be a shortage of supply. In this case, patients with indications will not be able to obtain effective drug treatment in a timely manner, which may delay the disease and pose a serious threat to the health and even life safety of patients. Summary of the invention

[0006] The purpose of the present invention is to provide a production line capacity intelligent management method for pharmaceutical production to solve the above technical problems.

[0007] The purpose of the present invention can be achieved through the following technical solutions: A production line capacity intelligent management method for drug production includes the following steps: Step S1: obtaining the indications of the drug, setting a historical data collection interval, and obtaining historical data, wherein the historical data includes the sales volume of the drug and the incidence rate of the indications; and obtaining a basic relationship between the sales volume and the indications according to the historical data; Step S2: Obtain an incidence curve based on the historical data, and perform periodic analysis on the incidence curve to obtain a number of incidence cycles; obtain all period curves corresponding to the incidence cycle, and obtain year weight coefficients based on all period curves of the incidence cycle; and obtain a standard incidence curve of the incidence cycle based on all period curves and year weight coefficients; Step S3: When entering a new disease cycle, the new disease cycle is recorded as the current disease cycle; a real-time incidence curve of the current disease cycle is generated in real time, and the standard incidence curve and the standard incidence curve are compared in real time to obtain a predicted incidence; based on the predicted incidence and the basic relationship, the recommended yield of the drug is obtained.

[0008] As a further solution of the present invention: the process of setting the historical data collection interval includes: Set a year threshold Y, where the setting range of the year threshold is Y∈[3,5]; obtain the current year, and the historical data collection interval is the time period from the previous Y years to the current year; A time interval threshold is set, and the setting range of the time interval threshold is [1, 7] days. In the historical data collection interval, a node is selected every time interval threshold and recorded as a production node, so that a plurality of production nodes are obtained.

[0009] As a further solution of the present invention: the historical data includes historical sales records of drugs and historical incidence records of indications, the historical sales records are the sales volume of drugs at each production node within the historical data collection interval, and the historical incidence records are the incidence rates of indications at each production node within the historical data collection interval.

[0010] As a further solution of the present invention: the process of obtaining the basic relationship includes: Establish an initial regression model, input the historical incidence records and historical sales records into the initial regression model, train the initial regression model, and obtain a final regression model; through the final regression model, obtain the basic relationship between sales volume Sv and incidence rate Ir Sv=Ir×β reg ×β dose +β noi , where β reg is the regulating factor, βdose is the dosage for a single treatment, β noi is the noise factor.

[0011] As a further solution of the present invention: the process of obtaining the incidence rate curve includes: Obtain the incidence rate at each production node in the year, and obtain the incidence rate set {Ir1, Ir2, ..., Ir num}, where Ir num Represents the incidence rate at the numth production node in a year, and num represents the total number of production nodes in a year; based on the incidence rate set of each year, the mean incidence rate of the historical disease record is obtained , where Ir i Represents the incidence rate at the i-th production node in the r-th year, i∈[1, num] and i is a positive integer, r∈[1, Y] and r is a positive integer; and obtains the standard deviation of the historical incidence record ; Based on the mean incidence rate, the standard scores for each year were obtained. , where Z r Represents the standard score of the rth year; sets a standard score threshold. If the standard score in a year exceeds the standard score threshold, the incidence rate of that year is recorded as abnormal; otherwise, the incidence rate is not abnormal; screens and eliminates the years with abnormal incidence rates in the historical incidence records, and generates an incidence rate curve based on the incidence rates at each production node in the remaining years.

[0012] As a further solution of the present invention: the process of periodically analyzing the incidence curve includes: A periodic function model is established, wherein the periodic function model is based on a sine function, and the incidence curve is fitted to the periodic function model by the least squares method. The fitting error is minimized by adjusting the parameters of the periodic function model to obtain a periodic function expression of the incidence curve. According to the periodic function expression, the period of the incidence curve is obtained, and one year is divided into several incidence periods according to the period.

[0013] As a further solution of the present invention: the process of obtaining the year weight coefficient includes: Obtain the periodic curve corresponding to the disease cycle in each year, and obtain the set {I1(y), I2(y), ..., I m (y)}, where y is the number of the production node, I m (y) represents the periodic curve of the mth year, and m is the total number of incidence curves; the elements in the set are randomly combined in pairs to obtain a number of combinations; For any combination, the combination is recorded as [I w1 (y), Iw2 (y)], where w1∈[1, m], w2∈[1, m], w1 and w2 are both positive integers, and w1≠w2; then the year difference value Ydv=|w1-w2| is obtained, and the curve difference value is obtained , where t1 is the number of the first production node in the outbreak period, t2 is the number of the last production node in the outbreak period, and y k represents the kth production node in the outbreak period, k∈[t1, t2] and k is a positive integer; According to the year difference value and curve difference value of each combination, the year weight coefficient is obtained , where h is the total number of combinations, Cdv e Indicates the curve difference value of the e-th combination, Ydv e represents the year difference value of the e-th combination, and λ is the preset correction coefficient.

[0014] As a further solution of the present invention: the process of obtaining the standard incidence rate curve includes: According to the year weight coefficient, the standard value of the incidence rate of each cycle curve of the disease cycle at the same production node is obtained. , where w g represents the year corresponding to the incidence curve of the g-th cycle curve, w0 is the current year, Ir g ´ represents the incidence rate of the g-th cycle curve at the production node; the standard incidence rate curve is generated according to the number of each production node in the disease cycle and its corresponding standard value of incidence rate.

[0015] As a further solution of the present invention: the process of obtaining the recommended yield includes: Obtain the difference between the standard incidence rate curve and the real-time incidence rate curve in real time , where G(y p ) represents the expression of the standard incidence curve, where G´(y p ) represents the expression of the real-time incidence curve, y p represents the pth production node, t p Indicates the number of the latest production node; Set a difference threshold. If the difference value is less than or equal to the difference threshold, the predicted incidence rate is the standard value of the incidence rate of the next production node. If the difference value exceeds the difference threshold, the predicted incidence rate PIr=G´(t p +1)+Dv, where G´(t p +1) indicates the incidence rate of the next production node in the real-time incidence rate curve; The recommended yield , where PIr sIt represents the predicted incidence rate of the sth indication of the drug at the next production node, and M is the total number of indications of the drug.

[0016] Beneficial effects of the present invention: The present invention directly relates the incidence rate to the drug sales volume through a regression model, combines the drug dosage and noise correction, accurately quantifies the real demand, avoids the deviation of traditional empirical prediction, and reduces the risk of overproduction or insufficient supply; and based on the comparison between the real-time incidence rate curve and the standard curve, quickly corrects the prediction deviation, adapts to sudden fluctuations of the disease (such as infectious disease outbreaks), and realizes dynamic matching of production and demand; the generation of the standard incidence rate curve eliminates abnormal years, integrates the historical cycle curve (weighted year weight coefficient), extracts the periodic law of the disease (such as seasonal influenza), provides a stable benchmark for production planning, reduces the interference of random fluctuations, and supports long-term capacity allocation; through the analysis of year differences and curve differences, the data of different years are given differentiated weights to enhance the generation of the standard curve The data can be represented and resisted by interference in order to maximize the value of historical data; by predicting the incidence rate of the next production node, the output can be adjusted in advance to avoid inventory waste caused by a sudden drop in demand and reduce inventory backlogs; the incidence trend can be monitored in real time, and production capacity can be increased before a surge in demand to ensure the accessibility of drugs (such as the rapid increase in production of antiviral drugs in the early stages of the epidemic) and prevent supply shortages; through accurate prediction, the company's production resources can be optimized (such as raw material procurement, production line scheduling), the overall operating cost can be reduced, and the social emergency response capability can be improved; the present invention replaces extensive production with data-driven, and through the closed-loop logic of "modeling quantified demand-periodic law extraction-real-time dynamic adjustment", the scientific, agile and low-cost control of drug production can be achieved, which has both commercial benefits and public health value. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 It is a flow chart of a method for intelligent management of production capacity of a production line for drug production according to the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] See also Figure 1 As shown, the present invention is a production line capacity intelligent management method for drug production, comprising the following steps: Step S1: set a historical data collection interval, select several production nodes in equal parts in the historical data collection interval; obtain all indications of the drug, obtain the historical sales record of the drug and the historical incidence record of the indication, the historical sales record is the sales volume of the drug at each production node within the historical data collection interval, and the historical incidence record is the incidence rate of the indication at each production node within the historical data collection interval; Establish an initial regression model, input the historical incidence records and historical sales records into the initial regression model, train the initial regression model, and obtain a final regression model; through the final regression model, obtain the basic relationship between sales volume Sv and incidence rate Ir Sv=Ir×β reg ×β dose +β noi , where β reg is the regulating factor, β dose is the dosage for a single treatment, β noi is the noise factor, and the units of the adjustment factor and the noise factor are consistent with the units of sales volume; It should be noted that a time period is set as the data collection range to ensure that the data covers sufficient historical information; several time points (production nodes) are equally selected within the collection interval to record drug sales and indication incidence rates; historical sales records (sales) of drugs and historical incidence records (incidence) of indications are collected to ensure that the data are aligned in the time dimension; regression analysis methods are used to take incidence rate (Ir) as the independent variable and drug sales volume (Sv) as the dependent variable to establish a mathematical relationship between the two; the model is trained through historical incidence records and sales records to optimize model parameters so that it can accurately fit historical data; after training, the relationship between sales volume and incidence rate is obtained, in which the adjustment factor is used to calibrate the final regression model so that the degree of influence of incidence rate on sales volume is adjusted, β dose Represents the dosage of the drug for each indication. The noise factor is used to correct the random errors or uncontrollable factors in the model, such as the impact of market share. The model is trained through historical data to reveal the potential pattern between incidence and sales. As a preferred embodiment of the present invention, a year threshold Y is set, and the setting range of the year threshold is Y∈[3,5]; the current year is obtained, and the historical data collection interval is the time period between the previous Y years and the current year; As a preferred embodiment of the present invention, the process of selecting the production node includes: A time interval threshold is set, and the setting range of the time interval threshold is [1, 7] days. In the historical data collection interval, a node is selected every time interval threshold, which is recorded as a production node, and a number of production nodes are obtained; It should be noted that historical data of 3 to 5 years can reflect recent trends while avoiding interference of too old data on the model; this time range usually contains enough data points to capture the cyclical or trend changes in incidence and sales; in addition, the time interval of 1 to 7 days can ensure data density while avoiding data being too sparse or redundant; adjust the time interval according to actual needs, for example: a shorter interval (1 day) is suitable for high-frequency data collection scenarios (such as acute infectious disease drugs); a longer interval (7 days) is suitable for low-frequency data collection scenarios (such as chronic disease drugs); through a reasonable time range and interval, ensure that the collected data is representative and timely; As a preferred embodiment of the present invention, the process of obtaining the incidence rate is as follows: Obtain the medical records of the medical institution, wherein the medical records include all patients and their symptoms; obtain the total number of patients according to the medical records, and obtain the number of patients whose symptoms are the indications, recorded as the number of indications, and obtain the incidence rate Ir=n / N, wherein n is the number of indications and N is the total number of patients; Step S2: According to the historical disease records, the years with abnormal incidence rates in the historical disease records are screened and eliminated, and the incidence rate curve is generated according to the incidence rates at each production node in the remaining years; the incidence rate curve of each year is analyzed for periodicity, a year is divided into several disease cycles, and the curve segments corresponding to the disease cycles on the incidence curve are recorded as periodic curves; For any disease cycle, obtain the cycle curve corresponding to the disease cycle in each year, and obtain the set {I1(y), I2(y), ..., I m (y)}, where y is the number of the production node, I m (y) represents the cycle curve of the mth year, where m is the total number of incidence curves; randomly combine the elements in the set in pairs to obtain several combinations, and obtain the year difference value and curve difference value of each combination; and obtain the year weight coefficient according to the year difference value and curve difference value of each combination; according to the cycle curve corresponding to the incidence cycle in each year and the year weight coefficient, obtain the standard incidence curve of the incidence cycle; It should be noted that by eliminating years with abnormal incidence rates, we can avoid the interference of abnormal data on model training and ensure the accuracy and reliability of the data. After eliminating outliers, the incidence curve can better reflect the real disease trend and provide a high-quality data basis for subsequent analysis. A year is divided into several disease cycles to capture the seasonal and cyclical characteristics of the disease (such as influenza is more prevalent in winter). By extracting curve segments of each disease cycle, we can refine and analyze the changing patterns of the disease in different time periods to support more accurate predictions. With the passage of time, disease epidemic patterns, environmental factors (such as climate, population density) and medical conditions may change. Recent data can more accurately reflect the current disease epidemic trends and drug demand. Therefore, by randomly combining two by two, analyzing the year difference value and curve difference value, we give different weights to data from different years to reflect their contribution. Combined with the year weight coefficient, we generate a standard incidence curve to eliminate random fluctuations and highlight the core trend of the disease. It is understandable that the standard incidence curve, as a benchmark, can be used for real-time comparison and calibration to improve the accuracy of predictions; the periodic analysis and standard curve generation methods are applicable to a variety of diseases (such as seasonal infectious diseases and chronic diseases) and have a wide range of application value; based on the standard incidence curve, companies can adjust drug production in advance to avoid insufficient supply or inventory backlogs; As a preferred embodiment of the present invention, the process of generating the incidence rate curve includes: Numbering each production node within a year, establishing a coordinate system with the number of the production node as the horizontal coordinate and the incidence rate as the vertical coordinate; converting the number of each production node and its corresponding incidence rate into coordinate points at corresponding positions on the coordinate system; and connecting each coordinate point with a smooth curve, recording the curve as an incidence rate curve; As a preferred embodiment of the present invention, the process of screening out the years with abnormal incidence rates in the historical incidence records includes: Obtain the incidence rate at each production node in the year, and obtain the incidence rate set {Ir1, Ir2, ..., Ir num}, where Ir num Represents the incidence rate at the numth production node in a year, and num represents the total number of production nodes in a year; based on the incidence rate set of each year, the mean incidence rate of the historical disease record is obtained , where Ir i Represents the incidence rate at the i-th production node in the r-th year, i∈[1, num] and i is a positive integer, r∈[1, Y] and r is a positive integer; and obtains the standard deviation of the historical incidence record ; Based on the mean incidence rate, obtain the standard score for each year , where Z rrepresents the standard score of the rth year; sets a standard score threshold, if the standard score in a year exceeds the standard score threshold, then the incidence rate of that year is abnormal; otherwise, the incidence rate is not abnormal; As a preferred embodiment of the present invention, the process of periodically analyzing the incidence rate curves of each year includes: Establishing a periodic function model, wherein the periodic function model is based on a sine function, fitting the incidence curve to the periodic function model by the least square method, and minimizing the fitting error by adjusting the parameters of the periodic function model to obtain a periodic function expression of the incidence curve; obtaining the period of the incidence curve according to the periodic function expression, and dividing a year into a number of incidence periods according to the period; As a preferred embodiment of the present invention, the process of obtaining the year difference value and the curve difference value includes: For any combination, the combination is recorded as [I w1 (y), I w2 (y)], where w1∈[1, m], w2∈[1, m], w1 and w2 are both positive integers, and w1≠w2; then the year difference value Ydv=|w1-w2| is obtained, and the curve difference value is obtained , where t1 is the number of the first production node in the outbreak period, t2 is the number of the last production node in the outbreak period, and y k represents the kth production node in the outbreak period, k∈[t1, t2] and k is a positive integer; As a preferred embodiment of the present invention, the process of obtaining the year weight coefficient includes: According to the year difference value and curve difference value of each combination, the year weight coefficient is obtained , where h is the total number of combinations, Cdv e Indicates the curve difference value of the e-th combination, Ydv e represents the year difference value of the e-th combination, and λ is the preset correction coefficient; As a preferred embodiment of the present invention, the process of obtaining the standard incidence rate curve of the disease cycle includes: According to the year weight coefficient, the standard value of the incidence rate of each cycle curve of the disease cycle at the same production node is obtained. , where w g represents the year corresponding to the incidence curve of the g-th cycle curve, w0 is the current year, Ir g ´ represents the incidence rate of the g-th cycle curve at the production node; the standard incidence rate curve is generated according to the number of each production node in the disease cycle and its corresponding standard value of incidence rate; Step S3: When entering a new disease cycle, the new disease cycle is recorded as the current disease cycle, the disease incidence rate at each production node of the current disease cycle is monitored in real time, a real-time disease incidence curve is generated, and a standard disease incidence curve of the current disease cycle is obtained; the standard disease incidence curve is compared with the standard disease incidence curve in real time, and the disease incidence rate of the next production node is predicted, which is recorded as the predicted disease incidence rate; according to the predicted disease incidence rate of each indication and the basic relationship, the recommended output of the drug at the next production node is obtained; It can be understood that during the current disease cycle, the disease incidence data is collected by production node to generate a real-time disease incidence curve, which reflects the latest disease epidemic trend; the standardized disease incidence curve generated based on historical data reflects the typical trend of the disease in a specific disease cycle; the standard curve is used as a benchmark for prediction and calibration to evaluate the deviation of real-time data; the real-time disease incidence curve is compared with the standard disease incidence curve, and the difference between the two is analyzed. Based on the comparison results, the disease incidence of the next production node is predicted (predicted disease incidence), which provides a basis for production adjustment; As a preferred embodiment of the present invention, the process of predicting the incidence rate includes: Obtain the difference between the standard incidence rate curve and the real-time incidence rate curve in real time , where G(y p ) represents the expression of the standard incidence curve, where G´(y p ) represents the expression of the real-time incidence curve, y p represents the pth production node, t p Indicates the number of the latest production node; Set a difference threshold. If the difference value is less than or equal to the difference threshold, the predicted incidence rate is the standard value of the incidence rate of the next production node. If the difference value exceeds the difference threshold, the predicted incidence rate PIr=G´(t p +1)+Dv, where G´(t p +1) indicates the incidence rate of the next production node on the real-time incidence rate curve; It is worth noting that in actual scenarios, there may be delays in drug production, distribution, and supply, resulting in some patients failing to obtain drugs in a timely manner; these unmet needs will accumulate and manifest as additional morbidity or drug demand at subsequent time points; therefore, the difference value Dv needs to be added to ensure that the needs of patients who fail to obtain drugs in a timely manner are met; As a preferred embodiment of the present invention, the process of obtaining the recommended yield includes: The recommended yield , where PIr s It represents the predicted incidence rate of the sth indication of the drug at the next production node, and M is the total number of indications of the drug.

[0021] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A production line capacity intelligent management method for drug production, characterized in that: The following steps are involved: Step S1: obtaining the indications of the drug, setting a historical data collection interval, and obtaining historical data, wherein the historical data includes the sales volume of the drug and the incidence rate of the indications; and obtaining a basic relationship between the sales volume and the indications according to the historical data; Step S2: Obtain an incidence curve based on the historical data, and perform periodic analysis on the incidence curve to obtain a number of incidence cycles; obtain all period curves corresponding to the incidence cycle, and obtain year weight coefficients based on all period curves of the incidence cycle; and obtain a standard incidence curve of the incidence cycle based on all period curves and year weight coefficients; Step S3: When entering a new disease cycle, the new disease cycle is recorded as the current disease cycle; a real-time incidence curve of the current disease cycle is generated in real time, and the standard incidence curve and the standard incidence curve are compared in real time to obtain a predicted incidence; based on the predicted incidence and the basic relationship, the recommended yield of the drug is obtained.

2. The method for intelligent management of production line capacity for drug production according to claim 1, characterized in that: In step S1, the process of setting the historical data collection interval includes: Set a year threshold Y, where the setting range of the year threshold is Y∈[3,5]; obtain the current year, and the historical data collection interval is the time period from the previous Y years to the current year; A time interval threshold is set, and the setting range of the time interval threshold is [1, 7] days. In the historical data collection interval, a node is selected every time interval threshold and recorded as a production node, so that a plurality of production nodes are obtained.

3. The method for intelligent management of production line capacity for drug production according to claim 2, characterized in that: In step S1, the historical data includes historical sales records of drugs and historical incidence records of indications. The historical sales records are the sales volume of drugs at each production node within the historical data collection interval, and the historical incidence records are the incidence rates of indications at each production node within the historical data collection interval.

4. The method for intelligent management of production line capacity for drug production according to claim 3, characterized in that: In step S1, the process of obtaining the basic relationship includes: Establish an initial regression model, input the historical incidence records and historical sales records into the initial regression model, train the initial regression model, and obtain a final regression model; through the final regression model, obtain the basic relationship between sales volume Sv and incidence rate Ir Sv=Ir×β reg ×β dose +β noi , where β reg is the regulating factor, β dose is the dosage for a single treatment, β noi is the noise factor.

5. The method for intelligent management of production line capacity for drug production according to claim 3, characterized in that: In step S2, the process of obtaining the morbidity curve includes: Obtain the incidence rate at each production node in the year, and obtain the incidence rate set {Ir1, Ir2, ..., Ir num }, where Ir num Represents the incidence rate at the numth production node in a year, and num represents the total number of production nodes in a year; based on the incidence rate set of each year, the mean incidence rate of the historical disease record is obtained , where Ir i Represents the incidence rate at the i-th production node in the r-th year, i∈[1, num] and i is a positive integer, r∈[1, Y] and r is a positive integer; and obtains the standard deviation of the historical incidence record ; Based on the mean incidence rate, the standard scores for each year were obtained. , where Z r Represents the standard score of the rth year; sets a standard score threshold. If the standard score in a year exceeds the standard score threshold, the incidence rate of that year is recorded as abnormal; otherwise, the incidence rate is not abnormal; screens and eliminates the years with abnormal incidence rates in the historical incidence records, and generates an incidence rate curve based on the incidence rates at each production node in the remaining years.

6. The method for intelligent management of production line capacity for drug production according to claim 1, characterized in that: In step S2, the process of periodically analyzing the incidence curve includes: A periodic function model is established, wherein the periodic function model is based on a sine function, and the incidence curve is fitted to the periodic function model by the least squares method. The fitting error is minimized by adjusting the parameters of the periodic function model to obtain a periodic function expression of the incidence curve. According to the periodic function expression, the period of the incidence curve is obtained, and one year is divided into several incidence periods according to the period.

7. The method for intelligent management of production line capacity for drug production according to claim 3, characterized in that: In step S2, the process of obtaining the year weight coefficient includes: Obtain the periodic curve corresponding to the disease cycle in each year, and obtain the set {I1(y), I2(y), ..., I m (y)}, where y is the number of the production node, I m (y) represents the periodic curve of the mth year, and m is the total number of incidence curves; the elements in the set are randomly combined in pairs to obtain a number of combinations; For any combination, the combination is recorded as [I w1 (y), I w2 (y)], where w1∈[1, m], w2∈[1, m], w1 and w2 are both positive integers, and w1≠w2; then the year difference value Ydv=|w1-w2| is obtained, and the curve difference value is obtained , where t1 is the number of the first production node in the outbreak period, t2 is the number of the last production node in the outbreak period, and y k represents the kth production node in the outbreak period, k∈[t1, t2] and k is a positive integer; According to the year difference value and curve difference value of each combination, the year weight coefficient is obtained , where h is the total number of combinations, Cdv e Indicates the curve difference value of the e-th combination, Ydv e represents the year difference value of the e-th combination, and λ is the preset correction coefficient.

8. The method for intelligent management of production line capacity for drug production according to claim 7, characterized in that: In step S2, the process of obtaining the standard incidence rate curve includes: According to the year weight coefficient, the standard value of the incidence rate of each cycle curve of the disease cycle at the same production node is obtained. , where w g represents the year corresponding to the incidence curve of the g-th cycle curve, w0 is the current year, Ir g ´ represents the incidence rate of the g-th cycle curve at the production node; the standard incidence rate curve is generated according to the number of each production node in the disease cycle and its corresponding standard value of incidence rate.

9. The method for intelligent management of production line capacity for drug production according to claim 7, characterized in that: In step S3, the process of obtaining the recommended output includes: Obtain the difference between the standard incidence rate curve and the real-time incidence rate curve in real time , where G(y p ) represents the expression of the standard incidence curve, where G´(y p ) represents the expression of the real-time incidence curve, y p represents the pth production node, t p Indicates the number of the latest production node; Set a difference threshold. If the difference value is less than or equal to the difference threshold, the predicted incidence rate is the standard value of the incidence rate of the next production node. If the difference value exceeds the difference threshold, the predicted incidence rate PIr=G´(t p +1)+Dv, where G´(t p +1) indicates the incidence rate of the next production node in the real-time incidence rate curve; The recommended yield , where PIr s It represents the predicted incidence rate of the sth indication of the drug at the next production node, and M is the total number of indications of the drug.

Images