Traditional Chinese medicine consistency mixed batch feeding method based on snow ablation algorithm
Through the snow ablation algorithm combined with the synchronous optimization of fingerprint map and index component content, the quality consistency of traditional Chinese medicine preparations was solved, the macroscopic and absolute content control of mixed batch feeds of traditional Chinese medicine was achieved, and the uniformity and safety of traditional Chinese medicine preparations were improved.
Patent Information
- Application Number
- CN202411057812.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-19
AI Technical Summary
The consistency of quality of traditional Chinese medicine preparations is difficult to guarantee, mainly due to the inconsistent active ingredients of different batches in different time periods and uneven content of medicinal materials, resulting in large differences in effective content, which affects the efficacy and safety.
Using the consistent mixed batch feeding method of traditional Chinese medicine based on snow ablation algorithm, the optimization objective function with the lowest mass square variance, absolute square variance and cost is constructed by combining the fingerprint map and index component content, and the chaotic strategy, exploration strategy and Levi's flight strategy carried out multi-target Pareto solution to obtain the mixing ratio of each batch of medicinal materials.
It realizes effective control of the quality of traditional Chinese medicine in terms of macroscopic and absolute content, improves the efficiency of mixed batch blending ratio, ensures high similarity with standard drugs, reduces the composition gap, and improves the uniformity and safety of traditional Chinese medicine preparations. sex.
Smart Images

Figure CN118982281A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of mixed batch blending of traditional Chinese medicines, and specifically provides a method for mixing and feeding traditional Chinese medicines in a consistent mixed batch based on a snow melting algorithm. Background Art
[0002] The uniformity and stability of the quality of traditional Chinese medicines directly affects the safety and effectiveness of patients' medication. However, due to many factors such as the diverse sources of medicinal materials, differences in pharmaceutical equipment and processes, and imperfect quality standards, the quality of the pharmaceutical process fluctuates and the consistency of product quality is difficult to ensure. At present, the uniformity of the quality of traditional Chinese medicines is difficult to control, mainly because the effective ingredients of different batches of the same variety and the same origin at different time periods are inconsistent, and the content of the effective ingredients of the same medicinal material is not uniform, resulting in large differences in the effective content of traditional Chinese medicine preparations.
[0003] Since the quality consistency of traditional Chinese medicine is directly related to clinical efficacy and safety, the application of mixed batch blending technology of traditional Chinese medicine has been highly valued. Since traditional Chinese medicine originates from nature, it is affected by many natural factors, such as origin, cultivation and processing technology, climate, etc. These factors may lead to differences in quality consistency. In recent years, researchers of traditional Chinese medicine have borrowed the blending technology of brewing technology to try to use mixed batch blending technology to stabilize the quality of traditional Chinese medicine, that is, to re-combine different batches of medicinal materials according to a certain proportion and then adjust and control the proportion of various components in the medicinal materials. The current blending research is mainly based on the chromatographic fingerprint of traditional Chinese medicine. The main blending methods include linear programming mixed homogenization and nonlinear least squares method. The above methods either use the similarity of fingerprints as the target or the content of index components as the target to calculate the blending coefficient. Although similarity evaluation can control the quality of traditional Chinese medicine from an overall macroscopic perspective, the similarity emphasizes the proportional relationship of the components and does not effectively control the absolute content of the components; although the content of index components can reflect the quality of traditional Chinese medicine to a large extent, it lacks overall macroscopic control of traditional Chinese medicine. Summary of the invention
[0004] In view of this, the object of the present invention is to provide a method for mixing and batching Chinese medicine consistency based on a snow melting algorithm, which can control important qualities in terms of macroscopic properties and absolute content.
[0005] By combining the fingerprint spectrum with the content of the index components for simultaneous optimization, the consistency of the mixed batch of traditional Chinese medicine can be ensured.
[0006] In order to achieve the above object, the present invention provides the following technical solutions:
[0007] A method for mixing and batching traditional Chinese medicines based on a snow melting algorithm comprises the following steps:
[0008] Step 1: Construct optimization objective function
[0009] The optimization objective function is to minimize the square difference of quality, the absolute square difference and the cost:
[0010]
[0011] Where: f1, f2 and f3 are the quality square difference objective function, absolute square difference objective function and cost objective function respectively; C mi represents the i-th peak area of the mixed drug; C si represents the i-th peak area of the standard drug; r represents the number of peak areas; p j represents the cost of the jth batch of medicinal materials; x j represents the blending coefficient of the jth batch of medicinal materials; t represents the number of batches of medicinal materials;
[0012] Step 2: Solve the objective function using the snow melting algorithm
[0013] 21) Using the chaos strategy to initialize the population, the individual positions are mapped to the squared mass difference, the absolute squared difference and the cost;
[0014] 22) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution;
[0015] 23) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels;
[0016] 24) The population is randomly divided into two sub-populations, the number of individuals in the two sub-populations is equal; and in the two sub-populations, one sub-population implements the exploration strategy and updates the individual positions; the other sub-population implements the development strategy and updates the individual positions;
[0017] 25) The individual positions in the two sub-populations are updated again using the Levy flight strategy;
[0018] 26) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution;
[0019] 27) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels;
[0020] 28) Determine whether the current optimal solution is better than the historical optimal solution: if so, replace the historical optimal solution with the current optimal solution; if not, keep the historical optimal solution unchanged;
[0021] 29) Determine whether the current number of iterations k is less than the set maximum number of iterations K: If so, set k=k+1 and execute step 24); if not, execute step 30);
[0022] 30) The objective function values of the mass square difference, absolute square difference and cost are obtained by mapping the individual positions of the optimal solution, and the blending ratio of each batch of medicinal materials is obtained.
[0023] Furthermore, the chaos strategy is expressed as:
[0024]
[0025] Where: x i represents the new population position; x i-1 Represents the original population position; u represents a random number from 0 to 1.
[0026] Further, the method of executing the exploration strategy to update the individual position is:
[0027]
[0028] The method to execute the development strategy and update the individual positions is:
[0029]
[0030] Where: Z i (t+1) represents the updated position of the i-th individual; Z i (t) represents the position of the i-th individual before the update; G(t) represents the position of the individual with the current optimal solution; Elite(t) represents an individual randomly selected from several elite groups in the population; BM i (t) represents the Brownian motion of the i-th individual; represents the centroid of the population; M represents the snowmelt rate; θ1 represents a random number from 0 to 1; θ2 represents a random number from 0 to 1; and:
[0031] The individuals Elite(t) randomly selected from several elite groups in the population are expressed as:
[0032] Elite(t)∈[G(t), Z second (t), Z third (t), Z C (t)]
[0033] Where: Z second (t) and Z third (t) respectively represent the second best individual position and the third best individual position in the current population; Z C (t) represents the centroid position of the top 50% of individuals: and:
[0034]
[0035] Where: N1 represents the number of individuals in the top 50%;
[0036] Brownian motion is expressed as:
[0037]
[0038] Among them: x represents the original population position;
[0039] The snowmelt rate M is expressed as:
[0040]
[0041] Where: t represents the number of iterations; t max Indicates the maximum number of iterations;
[0042] The centroid of the population It is expressed as:
[0043]
[0044] Where: N represents the number of individuals in the population.
[0045] Further, in step 25), the method of re-updating the individual position using the Levy flight strategy is:
[0046]
[0047] Where: Z i (t) represents the position of the i-th individual; is a point-to-point multiplication; l represents the step size control amount: Levy(λ) is a random search path and obeys the Levy distribution:
[0048] Levy~u=t (-λ) , 1<λ≤3
[0049] Levy flight is essentially a random step size, which is simulated using the Mantegma algorithm. The step size s is:
[0050]
[0051] Where: μ and v both represent normal distribution; the value range of β is 1<β<3; and:
[0052]
[0053] Where: μ Represents a constant that satisfies the following expression; σ v represents a constant whose value is 1; and:
[0054]
[0055] Where: Γ(·) represents the Gamma function.
[0056] The beneficial effects of the present invention are:
[0057] The consistent mixed batch feeding method of traditional Chinese medicine based on the snow melting algorithm of the present invention takes the minimum square difference of mass, the minimum absolute square difference and the lowest cost as the optimization goals, and then optimizes multiple objective functions by adopting the post-snow melting algorithm. It is suitable for solving the blending ratio under complicated conditions of traditional Chinese medicine blending, can improve the efficiency of the mixed batch blending ratio of traditional Chinese medicine, achieve the minimum difference with the effective ingredients of standard medicines, and the blending with high similarity with standard medicines, and can control important quality in terms of macroscopicity and absolute content. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:
[0059] Figure 1 It is a flow chart of the method for mixing and batching consistent traditional Chinese medicine based on the snow melting algorithm of the present invention;
[0060] Figure 2 Graphs of the random walk path and levy flight path. DETAILED DESCRIPTION
[0061] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.
[0062] like Figure 1 As shown, the method for mixing and feeding Chinese medicines in a consistent batch based on the snow melting algorithm in this embodiment includes the following steps:
[0063] Step 1: Construct optimization objective function
[0064] The optimization objective function is to minimize the square difference of quality, the absolute square difference and the cost:
[0065]
[0066] Where: f1, f2 and f3 are the quality square difference objective function, absolute square difference objective function and cost objective function respectively; C mi represents the i-th peak area of the mixed drug; C si represents the i-th peak area of the standard drug; r represents the number of peak areas; p j represents the cost of the jth batch of medicinal materials; x j represents the blending coefficient of the jth batch of medicinal materials; t represents the number of batches of medicinal materials.
[0067] Step 2: Solve the objective function using the snow melting algorithm
[0068] 21) A chaotic strategy is used to initialize the population and map individual positions into mass square difference, absolute square difference and cost.
[0069] In this embodiment, the chaos strategy is expressed as:
[0070]
[0071] Where: x i represents the new population position; x i -1 represents the original population position; u represents a random number from 0 to 1.
[0072] 22) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution.
[0073] 23) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels.
[0074] 24) The population is randomly divided into two sub-populations, and the number of individuals in the two sub-populations is equal. Among the two sub-populations, one sub-population implements the exploration strategy and updates the individual positions; the other sub-population implements the development strategy and updates the individual positions.
[0075] Specifically, the method of executing the exploration strategy to update the individual position is:
[0076]
[0077] The method to execute the development strategy and update the individual positions is:
[0078]
[0079] Where: Z i (t+1) represents the updated position of the i-th individual; Z i (t) represents the position of the i-th individual before the update; G(t) represents the position of the individual with the current optimal solution; Elite(t) represents an individual randomly selected from several elite groups in the population; BM i (t) represents the Brownian motion of the i-th individual; represents the centroid of the population; M represents the snowmelt rate; θ1 represents a random number from 0 to 1; θ2 represents a random number from 0 to 1.
[0080] The individuals Elite(t) randomly selected from several elite groups in the population are expressed as:
[0081] Elite(t)∈[G(t), Z second (t), Z third (t), Z c (t)]
[0082] Where: Z second(t) and Z third (t) respectively represent the second best individual position and the third best individual position in the current population; Z C (t) represents the centroid position of the top 50% of individuals: and:
[0083]
[0084] Where: N1 represents the number of individuals in the top 50%;
[0085] Brownian motion is expressed as:
[0086]
[0087] Where: x represents the original population position.
[0088] The snowmelt rate M is expressed as:
[0089]
[0090] Where: t represents the number of iterations; t max Indicates the maximum number of iterations.
[0091] The centroid of the population It is expressed as:
[0092]
[0093] Where: N represents the number of individuals in the population.
[0094] 25) The Levy flight strategy is used to update the individual positions in the two sub-populations again.
[0095] Lévy distribution describes a statistical physical phenomenon that contains long-tail characteristics. Based on this distribution, Lévy flight is defined as a random search path that combines short-distance searches with occasional long-distance searches. This pattern has been observed in nature and conforms to the foraging behavior trajectories of many animals and insects, such as bees, albatrosses, and fruit flies, and can explain natural phenomena such as Brownian motion. Lévy flight is valued for its potential in optimization algorithms, which can enhance the diversity of the population, expand the search range, and help the algorithm to more easily jump out of the local optimal solution, thereby improving the global search ability. In the field of evolutionary algorithms and intelligent optimization, such as cuckoo algorithm and particle swarm optimization algorithm, Lévy flight has been used to improve algorithm performance. The position update formula of Lévy flight usually includes a random step size that conforms to the Lévy distribution.
[0096] Since it is relatively complex to directly implement the Levy distribution, the Mantegna algorithm is often used for simulation. The Mantegna algorithm uses normal distribution and specific mathematical transformations to approximate the characteristics of the Levy distribution. When dealing with search problems in uncertain environments, the optimization algorithm using the Levy flight can more effectively expand the search range and improve the global optimization performance.
[0097] Specifically, the method of re-updating the individual position using the Levy flight strategy is:
[0098]
[0099] Where: Z i (t) represents the position of the i-th individual; is a point-to-point multiplication; l represents the step size control amount: Levy(λ) is a random search path and obeys the Levy distribution:
[0100] Levy~u=t (-λ) , 1<λ≤3
[0101] Levy flight is essentially a random step size, which is simulated using the Mantegma algorithm. The step size s is:
[0102]
[0103] Where: μ and v both represent normal distribution; the value range of β is 1<β<3; and:
[0104]
[0105]
[0106] Where: μ Represents a constant that satisfies the following expression; σ v represents a constant whose value is 1; and:
[0107]
[0108] Where: Γ(·) represents the Gamma function.
[0109] In order to illustrate the superiority of Levy flight, the Levy flight and random walk were simulated in Matlab, with a step length of 200 steps. The results are shown in Figure 2. Figure 2 As shown, it can be proved that Levy flight has a wider search capability. When using the optimization algorithm of Levy flight to deal with search problems in uncertain environments, it can more effectively expand the search range and improve the global optimization performance.
[0110] 26) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution.
[0111] 27) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels.
[0112] 28) Determine whether the current optimal solution is better than the historical optimal solution: if so, replace the historical optimal solution with the current optimal solution; if not, keep the historical optimal solution unchanged.
[0113] 29) Determine whether the current number of iterations k is less than the set maximum number of iterations K: If so, set k=k+1 and execute step 24); if not, execute step 30).
[0114] 30) The objective function values of the mass square difference, absolute square difference and cost are obtained by mapping the individual positions of the optimal solution, and the blending ratio of each batch of medicinal materials is obtained.
[0115] The specific implementation of this embodiment is described below with reference to specific examples.
[0116] This embodiment takes the mixed batch feeding of Radix Isatidis as an example, and optimizes the mixed batch feeding ratio by using the genetic optimization algorithm and the snow melting algorithm respectively.
[0117] The optimization results obtained by genetic optimization algorithm are shown in Table 1 and Table 2.
[0118] Table 1 Objective function and results of genetic algorithm for blending Radix Isatidis
[0119] Label Mass square deviation Blending Cost 1 1.6978E-003 83.9502 2 1.3578E-003 84.6032 3 1.2974E-003 84.6728 4 8.5637E-005 88.1394 5 7.8740E-005 88.2946 6 6.2359E-005 89.2946
[0120] Table 2 (2) Blending coefficients of Radix Isatidis using genetic algorithm
[0121] <![CDATA[x1]]> <![CDATA[x2]]> <![CDATA[x3]]> <![CDATA[x4]]> <![CDATA[x5]]> <![CDATA[x6]]> <![CDATA[x7]]> <![CDATA[x8]]> <![CDATA[x9]]> <![CDATA[x 10 ]]> 1 0.0580 0.0935 0.1714 0.1316 0.1435 0.0002 0.0446 0.1253 0.0429 0.1792 2 0.0883 0.0932 0.1410 0.1315 0.1433 0.0002 0.0447 0.1253 0.0436 0.1790 3 0.1614 0.0093 0.1277 0.1146 0.1844 0.0016 0 0.1253 0.0053 0.1152 4 0.0072 0.0028 0.1598 0.1347 0.1010 0.0027 0.1305 0.1265 0.0079 0.1381 5 0.0076 0.0027 0.1654 0.1306 0.0099 0.0027 0.1295 0.1265 0.0079 0.1375 6 0.1591 0.0095 0.1261 0.0114 0.0185 0.0016 0 0.0125 0.0053 0.0115
[0122] The optimization results obtained by the snow melting algorithm are shown in Tables 3 and 4.
[0123] Table 3 Objective function and results of Isatis root allocation in snow melting algorithm
[0124] Mass square deviation Absolute mean deviation price 1 7.57E-08 5.05E-04 92.3813 2 2.43E-07 9.68E-04 92.3828 3 1.82E-07 8.22E-04 92.3806 4 8.23E-08 5.66E-04 92.3915 5 1.46E-07 7.21E-04 92.3752 6 5.20E-08 4.27E-04 92.3672
[0125] Table 4 Blending coefficients of Radix Isatidis for snow melting algorithm
[0126] x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 1 0.0973 0.1336 0.1110 0.0694 0.0758 0.0860 0.0830 0.1324 0.1231 0.0812 2 0.0972 0.1338 0.1110 0.0695 0.0757 0.0858 0.0832 0.1324 0.1232 0.0813 3 0.0972 0.1338 0.1109 0.0695 0.0758 0.0858 0.0832 0.1324 0.1231 0.0813 4 0.0972 0.1337 0.1110 0.0694 0.0758 0.0859 0.0831 0.1324 0.1230 0.0813 5 0.0972 0.1338 0.1110 0.0695 0.0758 0.0857 0.0832 0.1324 0.1231 0.0813 6 0.0972 0.1338 0.1109 0.0694 0.0758 0.0858 0.0832 0.1323 0.1231 0.0812
[0127] By comparison, it can be seen that the improved multi-objective snow melting algorithm has more optimized objective functions and stronger constraints than the genetic algorithm, and the optimized results are more convergent, larger in magnitude, and more accurate. Compared with the genetic algorithm, the blending coefficient is more convergent, and every batch of medicinal materials can be used.
[0128] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.
Claims
1. A method for mixing and feeding Chinese medicines in a consistent batch based on a snow melting algorithm, characterized in that: The steps include: Step 1: Construct optimization objective function The optimization objective function is to minimize the square difference of quality, the absolute square difference and the cost: Where: f1, f2 and f3 are the quality square difference objective function, absolute square difference objective function and cost objective function respectively; C mi represents the i-th peak area of the mixed drug; C si represents the i-th peak area of the standard drug; r represents the number of peak areas; p j represents the cost of the jth batch of medicinal materials; x j represents the blending coefficient of the jth batch of medicinal materials; t represents the number of batches of medicinal materials; Step 2: Solve the objective function using the snow melting algorithm 21) Using the chaos strategy to initialize the population, the individual positions are mapped to the squared mass difference, the absolute squared difference and the cost; 22) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution; 23) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels; 24) The population is randomly divided into two sub-populations, the number of individuals in the two sub-populations is equal; and in the two sub-populations, one sub-population implements the exploration strategy and updates the individual positions; the other sub-population implements the development strategy and updates the individual positions; 25) The individual positions in the two sub-populations are updated again using the Levy flight strategy; 26) Calculate the objective function value corresponding to the individual position and calculate the multi-objective Pareto solution; 27) Perform non-dominated sorting based on the Pareto solution and record the positions of individuals with low dominance levels; 28) Determine whether the current optimal solution is better than the historical optimal solution: if so, replace the historical optimal solution with the current optimal solution; if not, keep the historical optimal solution unchanged; 29) Determine whether the current number of iterations k is less than the set maximum number of iterations K: If so, set k=k+1 and execute step 24); if not, execute step 30); 30) The objective function values of the mass square difference, absolute square difference and cost are obtained by mapping the individual positions of the optimal solution, and the blending ratio of each batch of medicinal materials is obtained.
2. The method for mixing and feeding Chinese medicines in a consistent batch based on the snow melting algorithm according to claim 1, characterized in that: The chaos strategy is expressed as: Where: x i represents the new population position; x i―1 represents the position of the original population; u represents a random number between 0 and 1.
3. The method for mixing and feeding Chinese medicines in a consistent batch based on the snow melting algorithm according to claim 1, characterized in that: The method to execute the exploration strategy to update the individual position is: The method to execute the development strategy and update the individual positions is: Where: Z i (t+1) represents the updated position of the i-th individual; Z i (t) represents the position of the i-th individual before the update; G(t) represents the position of the individual with the current optimal solution; Elite(t) represents an individual randomly selected from several elite groups in the population; BM i (t) represents the Brownian motion of the i-th individual; represents the centroid of the population; M represents the snowmelt rate; θ1 represents a random number between 0 and 1; θ2 represents a random number between 0 and 1; and: The individuals Elite(t) randomly selected from several elite groups in the population are expressed as: Elite(t)∈[G(t),Z second (t),Z third (t),Z C (t)] Where: Z second (t) and Z third (t) respectively represent the second best individual position and the third best individual position in the current population; Z C (t) represents the centroid position of the top 50% of individuals: and: Where: N1 represents the number of individuals in the top 50%; Brownian motion is expressed as: Among them: x represents the original population position; The snowmelt rate M is expressed as: Where: t represents the current iteration number; t max Indicates the maximum number of iterations; The centroid of the population It is expressed as: Where: N represents the number of individuals in the population.
4. The method for mixing and feeding Chinese medicines in a consistent batch based on the snow melting algorithm according to claim 1, characterized in that: In step 25), the method of re-updating the individual position using the Levy flight strategy is: Where: Z i (t) represents the position of the i-th individual; is a point-to-point multiplication; l represents the step size control amount: Levy(λ) is a random search path and obeys the Levy distribution: Levy~u=t (―λ) ,1<λ≤3 Levy flight is essentially a random step size, which is simulated using the Mantegma algorithm. The step size s is: Where: μ and ν both represent normal distribution; the value range of β is 1<β<3; and: Where: μ Represents a constant that satisfies the following expression; σ ν represents a constant whose value is 1; and: Where: Γ(·) represents the Gamma function.
