Pharmaceutical airflow pulse mixing system and method based on improved ant colony algorithm

Abstract

The invention discloses a pharmaceutical airflow pulse mixing system and method based on an improved ant colony algorithm. The method comprises the following steps: S1, forming a pharmaceutical micro-nano particle airflow pulse mixing standardized data set; s2, establishing a pharmaceutical micro-nano particle airflow pulse mixing process optimization model based on the pharmaceutical micro-nano particle airflow pulse mixing standardized data set; s3, outputting global optimization control parameters; s4, forming local optimization control parameter output; s5, according to the global optimization control parameter output and the local optimization control parameter output, performing real-time dynamic adjustment on the airflow pulse control parameter to form a closed-loop control strategy with the global control parameter and the local control parameter synergistic effect; and S6, applying the airflow pulse control parameters subjected to real-time dynamic adjustment to a pharmaceutical micro-nano particle airflow pulse mixing system, and carrying out mixing. Temperature parameters are dynamically adjusted through a closed-loop feedback mechanism so that the temperature parameters can meet the optimal mixing state, and it is ensured that particles are distributed more evenly in airflow.

Description

Technical Field

[0001] The present invention relates to the field of pharmaceutical technology, and particularly to a pharmaceutical air flow pulse mixing system and method based on an improved ant colony algorithm. Background Art

[0002] With the continuous development of the pharmaceutical industry, micro-nano particle preparations play an increasingly important role in drug delivery systems. Due to their high specific surface area, good solubility, and targeted delivery ability, micro-nano particles are widely used in controlled release preparations, nano-drug carriers, and pulmonary inhalation drug delivery fields. In these application scenarios, the mixing uniformity of particles plays a crucial role in the efficacy stability, bioavailability, and repeatability of the preparation process.

[0003] Currently, traditional pharmaceutical micro-nano particle air flow mixing methods mainly rely on fixed air flow parameter settings or empirical adjustment strategies, that is, setting the pressure, flow rate, and frequency parameters of air flow pulses according to experimental experience. However, the empirical method has significant defects: since the transport behavior of particles in the air flow is affected by the coupling of multiple factors, single parameter adjustment is difficult to meet the complex mixing requirements of different particle characteristics, often resulting in uneven distribution of particles in the flow field, local particle concentration being too high or agglomeration phenomena, affecting the quality stability of the final product.

[0004] On the other hand, in recent years, the application of intelligent optimization algorithms in pharmaceutical engineering has gradually increased. Existing research has attempted to use intelligent optimization methods such as ant colony algorithm and particle swarm optimization algorithm to optimize air flow pulse parameters to improve the mixing uniformity of particles. However, the existing optimization methods still have limitations:

[0005] Traditional intelligent optimization algorithms usually focus on global search or local search. In the optimization of micro-nano particle mixing, it is necessary to optimize both global air flow parameters and local particle transport characteristics simultaneously. Using ACO or PSO alone often results in insufficient global search ability or insufficient local search accuracy, making the adaptability of the optimization results under different particle conditions poor.

[0006] Existing optimization algorithms mainly optimize based on the objective function, and fail to fully consider the influence of physical properties such as particle size, surface charge, and density of micro-nano particles during the mixing process. The generalization ability of the optimization model for different particle types is low.

[0007] During the mixing process of micro-nano particles, temperature has an important influence on the agglomeration behavior, flow characteristics, and drug stability of particles. Existing methods fail to optimize and adjust according to temperature changes, resulting in instability of the mixing effect under different environmental conditions.

[0008] In summary, the existing pharmaceutical micro-nano particle gas flow pulse mixing method has obvious deficiencies in the balance between global and local search, the adaptability of particle physical properties, the real-time dynamic adjustment ability, and the optimization of temperature effects, making it difficult to meet the high requirements of the modern pharmaceutical industry for the mixing uniformity and stability of micro-nano particles. Summary of the Invention

[0009] An object of the present invention is to propose a pharmaceutical gas flow pulse mixing system and method based on an improved ant colony algorithm. The present invention dynamically adjusts the temperature parameter through a closed-loop feedback mechanism to make it meet the optimal mixing state, ensuring that the particles are more evenly distributed in the gas flow.

[0010] A pharmaceutical gas flow pulse mixing method based on an improved ant colony algorithm according to an embodiment of the present invention includes the following steps:

[0011] S1. Collect the physical property parameters of pharmaceutical micro-nano particles, and at the same time collect the initial control parameters of the gas flow pulse to form a standardized data set for the gas flow pulse mixing of pharmaceutical micro-nano particles;

[0012] S2. Establish an optimization model for the gas flow pulse mixing process of pharmaceutical micro-nano particles based on the standardized data set for the gas flow pulse mixing of pharmaceutical micro-nano particles;

[0013] S3. Use the improved ant colony algorithm to perform a global search on the objective function of the optimization model for the gas flow pulse mixing process of pharmaceutical micro-nano particles, optimize the global control parameters of the gas flow pulse, and output the global optimized control parameters;

[0014] S4. On the basis of obtaining the output of the global optimized control parameters, use the chaotic fruit fly optimization algorithm to refine and optimize the local transport behavior of pharmaceutical micro-nano particles in the gas flow to form the output of local optimized control parameters;

[0015] S5. Dynamically adjust the gas flow pulse control parameters in real time according to the output of the global optimized control parameters and the output of the local optimized control parameters to form a closed-loop control strategy in which the global and local control parameters act synergistically;

[0016] S6. Apply the gas flow pulse control parameters after the real-time dynamic adjustment to the pharmaceutical micro-nano particle gas flow pulse mixing system for implementation of mixing.

[0017] Optionally, S1 includes the following steps:

[0018] S11. Collect the physical property parameters of pharmaceutical micro-nano particles and establish a data set D of the physical property parameters of pharmaceutical micro-nano particles p :

[0019]

[0020] where D pRepresents the data set of physical characteristic parameters of pharmaceutical micro-nano particles, N p is the number of samples of pharmaceutical micro-nano particles collected, d p,i is the physical characteristic data of the i-th pharmaceutical micro-nano particle, δ i represents the particle size of the i-th pharmaceutical micro-nano particle, ρ i represents the density of the i-th pharmaceutical micro-nano particle, q i represents the surface charge of the i-th pharmaceutical micro-nano particle, ψ i represents the adhesion characteristic of the i-th pharmaceutical micro-nano particle;

[0021] S12. Collect the initial control parameters of the air flow pulse and establish the data set D of the initial control parameters of the air flow pulse f :

[0022]

[0023] Among them, D f represents the data set of the initial control parameters of the air flow pulse, N f is the number of samples of the air flow pulse parameters collected, d f,j is the initial control parameter of the j-th group of air flow pulses, P j represents the pressure of the j-th group of air flow pulses, f j represents the frequency of the j-th group of air flow pulses, T j represents the duration of the j-th group of air flow pulses, v j represents the flow velocity of the j-th group of air flow pulses;

[0024] S13. Correlate the data set D of the physical characteristic parameters of the pharmaceutical micro-nano particles p with the data set D of the initial control parameters of the air flow pulse f to form the mixed initial data set D of the pharmaceutical micro-nano particle air flow pulse m :

[0025] D m ={(d p,i ,d f,j )∣d p,i ∈D p ,d f,j ∈D f ,R(d p,i ,d f,j )}

[0026] Among them, D m represents the mixed initial data set of the pharmaceutical micro-nano particle air flow pulse, R(d p,i ,d f,j) is the correlation function between the physical properties of pharmaceutical micro-nano particles and the air flow pulse control parameters, constructed based on the influence relationships of particle size, particle density, surface charge, and adhesion properties on air flow pulse pressure, frequency, duration, and flow rate;

[0027] S14. Normalize the initial data set of pharmaceutical micro-nano particle air flow pulse mixing to generate a standardized data set of pharmaceutical micro-nano particle air flow pulse mixing And construct an index.

[0028] Optionally, the S2 includes the following steps:

[0029] S21. Based on the standardized data set of pharmaceutical micro-nano particle air flow pulse mixing Establish an optimization model M for the pharmaceutical micro-nano particle air flow pulse mixing process o :

[0030] M o =(X,Y,F);

[0031] Among them, X is the input parameter vector of the pharmaceutical micro-nano particle air flow pulse mixing process, defined as:

[0032]

[0033] Among them, is the particle size of the i-th pharmaceutical micro-nano particle after standardization, is the density of the i-th pharmaceutical micro-nano particle after standardization, is the surface charge of the i-th pharmaceutical micro-nano particle after standardization, is the adhesion property of the i-th pharmaceutical micro-nano particle after standardization, is the pressure of the j-th group of air flow pulses after standardization, is the frequency of the j-th group of air flow pulses after standardization, is the duration of the j-th group of air flow pulses after standardization, is the flow rate of the j-th group of air flow pulses after standardization;

[0034] S22. Define the target output parameter vector Y of the pharmaceutical micro-nano particle air flow pulse mixing:

[0035] Y=(U,C,E);

[0036] Among them, U represents the mixing uniformity index of pharmaceutical micro-nano particles, C represents the agglomeration rate index of pharmaceutical micro-nano particles, and E represents the energy consumption of the pharmaceutical micro-nano particle air flow pulse mixing process;

[0037] S23. Calculate the mixing uniformity index U of pharmaceutical micro-nano particles:

[0038]

[0039] Among them, N is the number of sampling points in the mixing region, and σ i is the standard deviation of particle concentration at the i-th sampling point, and is the average standard deviation of particle concentration in the overall mixing region;

[0040] Combined with the pharmaceutical micro-nano particle air flow pulse mixing standardized data set to analyze the influencing factors of the particle concentration standard deviation σ i The particle size affects the sedimentation rate of particles in the air flow, and thus affects the local concentration fluctuation. The particle density affects the drift behavior of particles in the pulsed air flow. The surface charge of the particles affects the electrostatic interaction between particles, resulting in local agglomeration, and thus affects the standard deviation of concentration. The particle adhesion characteristics affect the adhesion behavior of particles during the mixing process. The air flow pressure affects the transport intensity of particles. The air flow pulse frequency affects the periodic perturbation of the air flow on the particles, causing the particles to redistribute. The air flow pulse duration affects the residence time of particles in the mixing region, and thus changes the local concentration fluctuation. The air flow velocity affects the diffusion rate of particles;

[0041] S24. Calculate the pharmaceutical micro-nano particle agglomeration rate index C:

[0042]

[0043] Among them, M is the number of divisions in the particle agglomeration region, and n j represents the number of particles in the j-th agglomeration region, and N j represents the number of theoretically uniformly distributed particles in this region. n j is affected by the surface charge and adhesion characteristics of the pharmaceutical micro-nano particles;

[0044] S25. Calculate the energy consumption E of the pharmaceutical micro-nano particle air flow pulse mixing process:

[0045]

[0046] Among them, K is the number of times of air flow pulse action, is the pressure of the k-th air flow pulse after standardization, is the action time of the k-th air flow pulse after standardization; The energy consumption E is affected by the air flow pulse frequency and the flow velocity ;

[0047] S26. Construct the optimization objective function F(X) for the gas flow pulse mixing process of pharmaceutical micro-nano particles:

[0048] F(X) = w1U - w2C - w3E

[0049] Where w1, w2, and w3 are optimization weight coefficients;

[0050] S27. Simulate and calculate the mixing behavior of pharmaceutical micro-nano particles based on the optimization objective function F(X) of the gas flow pulse mixing process of pharmaceutical micro-nano particles, establish an optimization solution framework, and update the optimization model of the gas flow pulse mixing process of pharmaceutical micro-nano particles.

[0051] Optionally, step S3 includes the following steps:

[0052] S31. Define the optimization search space Ω according to the optimization objective function F(X) of the gas flow pulse mixing process of pharmaceutical micro-nano particles:

[0053]

[0054] Where is the normalized gas flow pulse pressure, is the normalized gas flow pulse frequency, is the normalized gas flow velocity;

[0055] S32. Adopt an adaptive pheromone initialization mechanism based on the input parameter vector X to construct an adaptive pheromone matrix τ based on the physical properties of pharmaceutical micro-nano particles and the gas flow pulse control parameters (0) :

[0056]

[0057] Where is the initial pheromone concentration at path (i, j), τ0 is the initial pheromone concentration reference value, indicating the pheromone level without optimization, is the surface charge of pharmaceutical micro-nano particle i after normalization, which affects the electrostatic interaction between particles, determines the particle aggregation trend, and thus affects the mixing uniformity U, is the mean value of the surface charge of pharmaceutical micro-nano particles, is the standard deviation of the surface charge of pharmaceutical micro-nano particles, is the mean value of the gas flow pulse pressure, is the standard deviation of the gas flow pulse pressure, is the mean value of the gas flow pulse velocity, is the standard deviation of the gas flow pulse velocity;

[0058] S33. Define the heuristic factor η at path (i, j) ij :

[0059]

[0060] Among them, The mean value of the sedimentation velocity, ξ ij Represents the sedimentation velocity of particles under the action of air flow pulses:

[0061]

[0062] Among them, Is the particle density, which affects the floating property of particles under the action of air flow pulses. The heuristic factor makes ants tend to choose paths that can improve the mixing uniformity U and reduce the particle agglomeration rate C when searching for paths;

[0063] S34. Calculate the transition probability of ant k in each iteration

[0064]

[0065] Among them, α and β respectively represent the pheromone intensity and the weight parameter of the heuristic factor, making the search tend to the direction of optimizing the objective function F(X);

[0066] S35. Combine the population behavior of micro-nano particles for dynamic pheromone update, so that the pheromone update is not only affected by the objective function, but also combines the surface charge and flow velocity Furthermore, optimize the search direction, and define the pheromone update as:

[0067]

[0068] Among them, Is the pheromone concentration at the path (i, j) when the ant is in the (t + 1)-th iteration, representing the preference degree of the ant for choosing this path. ρ is the pheromone evaporation coefficient, Is the pheromone concentration at the path (i, j) in the t-th iteration, inheriting the pheromone level of the previous round. m is the number of ants participating in the pheromone update, Is the pheromone increment released by the k-th ant on the path (i, j):

[0069]

[0070] Among them, Q is the total pheromone release coefficient, Is the fitness value of the k-th ant at the path (i, j), representing the contribution degree of this path to the optimization objective. The larger the fitness value, the better the path and the stronger the pheromone update;

[0071] S36. Optimize the search efficiency by using a dynamic evaporation coefficient:

[0072]

[0073] Among them, ρ min and ρ max are respectively the set minimum and maximum volatility coefficients, λ1 is the adjustment rate parameter, and t is the current iteration number;

[0074] S37. Output the global optimization control parameters:

[0075]

[0076] Optionally, the S4 includes the following steps:

[0077] S41. Construct a local optimization search space according to the global optimization control parameters, and define the optimization search space Ω ′ :

[0078]

[0079] Among them, represents the local optimization airflow pulse action angle, represents the local airflow disturbance frequency, represents the local airflow shear strength, represents the temperature of the local particle mixing area;

[0080] S42. Initialize the local optimization population by using the chaos fruit fly optimization algorithm based on the global optimization parameters, set the fruit fly population size to N f , and perform random initialization within the local optimization search space Ω ′ to make the initial individuals satisfy:

[0081]

[0082] Among them, is the individual position of the i-th fruit fly. By introducing the global optimization control parameters (P, f, v), the initial position of the fruit fly population is perturbed and adjusted to make its search direction consistent with the global optimization result:

[0083]

[0084] Among them, ∈1 is the local search perturbation factor, making the initial search of the fruit fly closer to the optimal solution;

[0085] S43. Calculate the local airflow mixing performance index of the fruit fly individual i at the position :

[0086]

[0087] Among them, Topt is the temperature regulation fitness, and w1, w2, w3, w4 are the weight coefficients;

[0088] S44. Use the fruit fly olfactory search mechanism for local optimization. Each fruit fly updates its current position according to its own air flow perturbation characteristics:

[0089]

[0090] where λ f is the local optimization step factor, is the current local optimal solution;

[0091] S45. Adopt a local search strategy based on chaotic perturbation. At each iteration, introduce a chaotic map for local perturbation:

[0092]

[0093] where μ(t) is the dynamic chaotic perturbation factor, is the current average position of all fruit fly individuals, and ξ is a random variable uniformly distributed between [0, 1];

[0094] S46. Combine the temperature change to optimize the particle mixing behavior and define the influence factor of temperature on particle mixing:

[0095]

[0096] where T eff is the actual acting temperature of the particles, is the temperature setting value of the current fruit fly individual, C is the particle agglomeration rate, and the higher the agglomeration rate, the greater the influence of temperature on mixing;

[0097] When the temperature stability condition is met, the current optimization state is acceptable; otherwise, perform temperature adjustment optimization:

[0098] |T eff -T| < ∈ T ;

[0099] where ∈ T is the temperature stability threshold;

[0100] S47. When the termination condition is met, output the optimal local mixing control parameters:

[0101]

[0102] where, is the optimal local air flow acting angle obtained by optimization, is the optimal local air flow perturbation frequency obtained by optimization, is the optimal local air flow shear strength obtained by optimization, The optimal local temperature control parameters obtained by optimization.

[0103] Optionally, the S5 includes the following steps:

[0104] S51. Based on the global optimization control parameter output and the local optimization control parameter output, construct an air flow pulse dynamic control model, and define the optimization variable vector X of the air flow pulse control parameter c :

[0105]

[0106] Construct an air flow pulse dynamic control model M c :

[0107] M c =(X c , F c )

[0108] where F c is the real-time dynamic control strategy function;

[0109] S52. Adopt a global and local collaborative control strategy to achieve adaptive adjustment of the air flow pulse, and define the dynamic adjustment function of the global and local collaboration control:

[0110]

[0111] where is the air flow pulse control parameter for the (t + 1)-th round of iteration, α g and α l are the global control weight and the local control weight respectively, ΔX g is the update increment of the global optimization control parameter, and ΔX l is the update increment of the local optimization control parameter;

[0112] S53. Optimize the control weights α g , α l using an adaptive weight adjustment strategy, and calculate the optimal control weights according to the real-time mixing uniformity U, the particle agglomeration rate C, and the energy consumption E:

[0113]

[0114] α l = 1 - α g

[0115] where w U , w C , w E are the adjustment weights of the mixing uniformity, the particle agglomeration rate, and the energy consumption respectively;

[0116] S54. Optimize the air flow pulse control strategy in combination with temperature changes, and define the adjustment function of the air flow pulse control parameters with respect to temperature:

[0117]

[0118] Among them, β T is the temperature adjustment coefficient, is the actual temperature of the current air flow pulse region, and T * is the optimal temperature parameter calculated by local optimization;

[0119] When the temperature deviation satisfies the stability condition:

[0120]

[0121] Among them, ∈ T is the temperature adjustment threshold, then stop adjusting the temperature control parameters, otherwise continue to optimize;

[0122] S55. Adopt a closed-loop feedback control strategy to adjust the air flow pulse parameters in real time. After each optimization iteration cycle t, calculate the error correction term:

[0123]

[0124] Among them, γ1 is the feedback adjustment gain, is the ideal air flow pulse control parameter;

[0125] Finally, update the control parameter:

[0126]

[0127] S56. When the optimization converges, output the optimal air flow pulse control parameters

[0128]

[0129] Among them, is the optimal air flow pulse pressure finally optimized, is the optimal air flow pulse frequency finally optimized, is the optimal air flow velocity finally optimized, is the optimal air flow action angle finally optimized, is the optimal air flow disturbance frequency finally optimized, is the optimal air flow shear strength finally optimized, is the optimal temperature control parameter finally optimized.

[0130] A pharmaceutical air flow pulse mixing system based on an improved ant colony algorithm is used to execute a pharmaceutical air flow pulse mixing method based on an improved ant colony algorithm.

[0131] The beneficial effects of the present invention are as follows:

[0132] (1) The present invention combines an improved ant colony algorithm and a chaotic fruit fly optimization algorithm to achieve dynamic coordination of global optimization and local optimization. In the global optimization stage, the improved ant colony algorithm optimizes the global control parameters of the air flow pulse through adaptive pheromone initialization, particle dynamics heuristic factor, and dynamic evaporation coefficient adjustment strategy, improving the convergence speed and global search ability of the optimization algorithm. In the local optimization stage, a fruit fly optimization algorithm based on chaotic perturbation is introduced. Using the chaotic mapping perturbation mechanism and local search adaptive step size control, the particle transport behavior is refined and optimized, enhancing the stability and accuracy of local search. By combining the two optimization algorithms, the optimization path has strong global exploration ability in the initial stage, and at the same time, the search accuracy is improved through local refinement optimization in the later stage, avoiding the search falling into local optimum and improving the effectiveness of air flow pulse parameter optimization.

[0133] (2) The present invention introduces an adaptive temperature control mechanism in the optimization process, and adjusts in real time for the temperature influencing factors in the air flow pulse mixing process, solving the problem that existing optimization methods ignore the influence of temperature changes on particle mixing behavior. In the hybrid optimization process, it is dynamically adjusted through the temperature influence factor combined with the particle agglomeration rate, so that the air flow pulse control parameters can adapt to different temperature changes, thereby improving the mixing stability of particles under different working conditions. The temperature-particle hybrid coupling optimization strategy is adopted, and the temperature parameters are dynamically adjusted through a closed-loop feedback mechanism to make it meet the optimal mixing state, ensuring that the distribution of particles in the air flow is more uniform.

[0134] (3) Through the dynamic coordination of global optimization control parameters and local optimization control parameters, the present invention constructs a closed-loop control strategy based on adaptive feedback adjustment, solving the problem that existing optimization methods lack real-time adjustment ability and cannot dynamically adjust parameters according to the particle mixing state. An adaptive weight adjustment mechanism is introduced in the optimization process. According to the mixing uniformity, particle agglomeration rate, and energy consumption monitored in real time, the optimal global and local optimization control weights are calculated to realize intelligent allocation of parameters, enabling the optimization algorithm to make adaptive adjustments according to real-time changes. Description of the Drawings

[0135] The drawings are used to provide further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, but do not constitute a limitation to the present invention. In the drawings:

[0136] Figure 1 is a flowchart of a pharmaceutical air flow pulse mixing system and method based on an improved ant colony algorithm proposed by the present invention. Detailed Embodiments

[0137] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only schematically showing the basic structure of the present invention, and therefore only showing the components related to the present invention.

[0138] Reference Figure 1 , a pharmaceutical gas flow pulse mixing method based on an improved ant colony algorithm, comprising the following steps:

[0139] S1. Collect the physical property parameters of pharmaceutical micro-nano particles, and at the same time collect the initial control parameters of the gas flow pulse to form a standardized data set for the gas flow pulse mixing of pharmaceutical micro-nano particles;

[0140] S2. Establish an optimization model for the gas flow pulse mixing process of pharmaceutical micro-nano particles based on the standardized data set for the gas flow pulse mixing of pharmaceutical micro-nano particles;

[0141] S3. Use the improved ant colony algorithm to perform a global search on the objective function of the optimization model for the gas flow pulse mixing process of pharmaceutical micro-nano particles, optimize the global control parameters of the gas flow pulse, and output the global optimized control parameters;

[0142] S4. On the basis of obtaining the output of the global optimized control parameters, use the chaotic fruit fly optimization algorithm to refine and optimize the local transport behavior of pharmaceutical micro-nano particles in the gas flow to form the output of local optimized control parameters;

[0143] S5. According to the output of the global optimized control parameters and the output of the local optimized control parameters, perform real-time dynamic adjustment on the gas flow pulse control parameters to form a closed-loop control strategy with the coordinated action of global and local control parameters;

[0144] S6. Apply the gas flow pulse control parameters after real-time dynamic adjustment to the gas flow pulse mixing system of pharmaceutical micro-nano particles to perform the mixing.

[0145] In this embodiment, S1 includes the following steps:

[0146] S11. Collect the physical property parameters of pharmaceutical micro-nano particles and establish a data set D of the physical property parameters of pharmaceutical micro-nano particles p :

[0147]

[0148] where D p represents the data set of the physical property parameters of pharmaceutical micro-nano particles, N p is the number of samples of pharmaceutical micro-nano particles collected, d p,i is the physical property data of the i-th pharmaceutical micro-nano particle, δ i represents the particle size of the i-th pharmaceutical micro-nano particle, ρ i represents the density of the i-th pharmaceutical micro-nano particle, qi Denote the surface charge of the i-th pharmaceutical micro-nano particle, ψ i Denote the adhesion property of the i-th pharmaceutical micro-nano particle;

[0149] S12. Collect the initial control parameters of the air flow pulse and establish the data set D of the initial control parameters of the air flow pulse f :

[0150]

[0151] where D f Denote the data set of the initial control parameters of the air flow pulse, N f is the number of samples of the air flow pulse parameters collected, d f,j is the j-th group of initial control parameters of the air flow pulse, P j Denote the pressure of the j-th group of air flow pulses, f j Denote the frequency of the j-th group of air flow pulses, T j Denote the duration of the j-th group of air flow pulses, v j Denote the flow velocity of the j-th group of air flow pulses;

[0152] S13. Correlate the data set D of the physical property parameters of the pharmaceutical micro-nano particles p with the data set D of the initial control parameters of the air flow pulse f to form the mixed initial data set D of the pharmaceutical micro-nano particle air flow pulse m :

[0153] D m ={(d p,i , d f,j )|d p,i ∈D p , d f,j ∈D f , R(d p,i , d f,j )}

[0154] where D m Denote the mixed initial data set of the pharmaceutical micro-nano particle air flow pulse, R(d p,i , d f,j ) is the correlation function between the physical properties of the pharmaceutical micro-nano particles and the control parameters of the air flow pulse, constructed based on the influence relationships of the particle size, particle density, surface charge and adhesion property on the air flow pulse pressure, frequency, duration and flow velocity;

[0155] S14. Normalize the mixed initial data set of the pharmaceutical micro-nano particle air flow pulse to generate the mixed standardized data set of the pharmaceutical micro-nano particle air flow pulse and construct an index.

[0156] In this embodiment, S2 includes the following steps:

[0157] S21. Based on the standardized data set of the pulsed gas flow mixing of pharmaceutical micro-nano particles Establish an optimization model M for the pulsed gas flow mixing process of pharmaceutical micro-nano particles o :

[0158] M o =(X, Y, F);

[0159] Among them, X is the input parameter vector of the pulsed gas flow mixing process of pharmaceutical micro-nano particles, defined as:

[0160]

[0161] Among them, is the particle size of the i-th standardized pharmaceutical micro-nano particle, is the density of the i-th standardized pharmaceutical micro-nano particle, is the surface charge of the i-th standardized pharmaceutical micro-nano particle, is the adhesion property of the i-th standardized pharmaceutical micro-nano particle, is the pressure of the j-th group of pulsed gas flows after standardization, is the frequency of the j-th group of pulsed gas flows after standardization, is the duration of the j-th group of pulsed gas flows after standardization, is the flow velocity of the j-th group of pulsed gas flows after standardization;

[0162] S22. Define the target output parameter vector Y of the pulsed gas flow mixing of pharmaceutical micro-nano particles:

[0163] Y=(U, C, E);

[0164] Among them, U represents the mixing uniformity index of pharmaceutical micro-nano particles, C represents the agglomeration rate index of pharmaceutical micro-nano particles, and E represents the energy consumption of the pulsed gas flow mixing process of pharmaceutical micro-nano particles;

[0165] S23. Calculate the mixing uniformity index U of pharmaceutical micro-nano particles:

[0166]

[0167] Among them, N is the number of sampling points in the mixing area, and σ i is the standard deviation of particle concentration at the i-th sampling point, is the average standard deviation of particle concentration in the overall mixing area;

[0168] Analyze the influencing factors of the standard deviation of particle concentration σ i in combination with the standardized data set of the pulsed gas flow mixing of pharmaceutical micro-nano particles, and the particle size Affect the sedimentation rate of particles in the air flow, and thus affect the local concentration fluctuation, particle density Affect the drift behavior of particles in the pulsed air flow, particle surface charge Affect the electrostatic interaction between particles, resulting in local agglomeration, and thus affect the standard deviation of concentration, particle adhesion characteristics Affect the adhesion behavior of particles during the mixing process, air flow pressure Affect the transport intensity of particles, air flow pulse frequency Affect the periodic perturbation of the air flow on the particles, causing the particles to redistribute, air flow pulse duration Affect the residence time of particles in the mixing area, and thus change the local concentration fluctuation, air flow velocity Affect the diffusion rate of particles;

[0169] S24. Calculate the agglomeration rate index C of pharmaceutical micro-nano particles:

[0170]

[0171] Among them, M is the number of divisions of the particle agglomeration area, n j represents the number of particles in the jth agglomeration area, N j represents the theoretically evenly distributed number of particles in this area, n j is affected by the surface charge and adhesion characteristics of pharmaceutical micro-nano particles;

[0172] S25. Calculate the energy consumption E of the pharmaceutical micro-nano particle air flow pulse mixing process:

[0173]

[0174] Among them, K is the number of times of air flow pulse action, is the pressure of the kth air flow pulse after standardization, is the action time of the kth air flow pulse after standardization; the energy consumption E is affected by the air flow pulse frequency and flow velocity ;

[0175] S26. Construct the optimization objective function F(X) for the pharmaceutical micro-nano particle air flow pulse mixing process:

[0176] F(X) = w1U - w2C - w3E

[0177] Among them, w1, w2, w3 are optimization weight coefficients;

[0178] S27. Simulate and calculate the mixing behavior of pharmaceutical micro-nano particles according to the optimization objective function F(X) of the air-pulse mixing process of pharmaceutical micro-nano particles, establish an optimization solution framework, and update the optimization model of the air-pulse mixing process of pharmaceutical micro-nano particles.

[0179] In this embodiment, S3 includes the following steps:

[0180] S31. Define the optimization search space Ω according to the optimization objective function F(X) of the air-pulse mixing process of pharmaceutical micro-nano particles:

[0181]

[0182] Among them, is the normalized air-pulse pressure, is the normalized air-pulse frequency, is the normalized air-flow velocity;

[0183] S32. Adopt an adaptive pheromone initialization mechanism based on the input parameter vector X to construct an adaptive pheromone matrix τ based on the physical properties of pharmaceutical micro-nano particles and air-pulse control parameters (0) :

[0184]

[0185] Among them, is the initial pheromone concentration at the path (i, j), τ0 is the initial pheromone concentration reference value, indicating the pheromone level without optimization, is the surface charge of the i-th pharmaceutical micro-nano particle after normalization, which affects the electrostatic interaction between particles, determines the particle agglomeration trend, and thus affects the mixing uniformity U, is the mean value of the surface charge of pharmaceutical micro-nano particles, is the standard deviation of the surface charge of pharmaceutical micro-nano particles, is the mean value of the air-pulse pressure, is the standard deviation of the air-pulse pressure, is the mean value of the air-pulse flow velocity, is the standard deviation of the air-pulse flow velocity;

[0186] S33. Define the heuristic factor η at the path (i, j) ij :

[0187]

[0188] Among them, The mean value of the sedimentation velocity, ξ ij represents the sedimentation velocity of the particles under the action of the air-pulse:

[0189]

[0190] Among them, is the particle density, which affects the floating property of the particles under the action of the air flow pulse. The inspiration factor makes the ants tend to choose the path that can improve the mixing uniformity U and reduce the particle agglomeration rate C when searching for the path;

[0191] S34. Calculate the transition probability of ant k in each iteration

[0192]

[0193] Among them, α and β respectively represent the pheromone intensity and the weight parameter of the inspiration factor, making the search tend to the direction of optimizing the objective function F(X);

[0194] S35. Combine the population behavior of micro-nano particles for dynamic pheromone update, so that the pheromone update is not only affected by the objective function, but also combines the surface charge and the flow velocity Furthermore, optimize the search direction, and define the pheromone update as:

[0195]

[0196] Among them, is the pheromone concentration at the path (i, j) when the (t + 1)-th iteration is carried out by the ant, indicating the preference degree selected by the ant on this path. ρ is the pheromone evaporation coefficient, is the pheromone concentration at the path (i, j) in the t-th iteration, inheriting the pheromone level of the previous round. m is the number of ants participating in the pheromone update, is the pheromone increment released by the k-th ant on the path (i, j):

[0197]

[0198] Among them, Q is the total pheromone release coefficient, is the fitness value of the k-th ant at the path (i, j), indicating the contribution degree of this path to the optimization objective. The larger the fitness value, the better the path and the stronger the pheromone update;

[0199] S36. Optimize the search efficiency by using a dynamic evaporation coefficient:

[0200]

[0201] Among them, ρ min and ρ max are respectively the set minimum and maximum evaporation coefficients, λ1 is the adjustment rate parameter, and t is the current iteration number;

[0202] S37. Output the globally optimized control parameters:

[0203]

[0204] In this embodiment, S4 includes the following steps:

[0205] S41. Construct a local optimization search space based on the globally optimized control parameters, and define the optimization search space Ω of the local hybrid control parameters ′ :

[0206]

[0207] Among them, represents the local optimized air flow pulse action angle, represents the local air flow disturbance frequency, represents the local air flow shear strength, represents the temperature of the local particle mixing area;

[0208] S42. Initialize the local optimization population using the chaotic fruit fly optimization algorithm based on the globally optimized parameters, set the fruit fly population size to N f , and perform random initialization within the local optimization search space Ω ′ so that the initial individuals satisfy:

[0209]

[0210] Among them, is the individual position of the i-th fruit fly. By introducing the globally optimized control parameters (P, f, v), the initial position of the fruit fly population is perturbed and adjusted to make its search direction consistent with the globally optimized result:

[0211]

[0212] Among them, ∈1 is the local search perturbation factor, making the initial search of the fruit fly closer to the optimal solution;

[0213] S43. Calculate the local air flow mixing performance index of the fruit fly individual i at the position :

[0214]

[0215] Among them, T opt is the temperature regulation fitness, and w1, w2, w3, w4 are weight coefficients;

[0216] S44. Perform local optimization using the fruit fly olfactory search mechanism. Each fruit fly updates its current position according to its own air flow disturbance characteristics:

[0217]

[0218] Among them, λ f is the local optimization step size factor, is the current local optimal solution;

[0219] S45. Adopt a local search strategy based on chaotic perturbation. At each iteration, introduce a chaotic map for local perturbation:

[0220]

[0221] Among them, μ(t) is the dynamic chaotic perturbation factor, is the current average position of all fruit fly individuals, ξ is a random variable uniformly distributed between [0, 1];

[0222] S46. Optimize the particle mixing behavior in combination with temperature changes, and define the influence factor of temperature on particle mixing:

[0223]

[0224] Among them, T eff is the actual acting temperature of the particles, is the temperature setting value of the current fruit fly individual, C is the particle agglomeration rate, and the higher the agglomeration rate, the greater the influence of temperature on mixing;

[0225] When the temperature stability condition is met, the current optimization state is acceptable, otherwise, perform temperature adjustment and optimization:

[0226] |T eff - T| < ∈ T ;

[0227] Among them, ∈ T is the temperature stability threshold;

[0228] S47. When the termination condition is met, output the optimal local mixing control parameters:

[0229]

[0230] Among them, is the optimal local air flow action angle obtained by optimization, is the optimal local air flow perturbation frequency obtained by optimization, is the optimal local air flow shear strength obtained by optimization, is the optimal local temperature control parameter obtained by optimization.

[0231] In this embodiment, S5 includes the following steps:

[0232] Output the global optimization control parameter output and the local optimization control parameter output, construct an air flow pulse dynamic control model, and define the optimization variable vector X of the air flow pulse control parameter c :

[0233]

[0234] Construct an air flow pulse dynamic control model M c :

[0235] M c =(X c ,F c )

[0236] where F c is a real-time dynamic control strategy function;

[0237] S52. Adopt a global and local collaborative control strategy to achieve adaptive adjustment of the air flow pulse, and define the dynamic adjustment function of global and local collaboration control:

[0238]

[0239] where is the air flow pulse control parameter in the (t + 1)-th iteration, α g and α l are the global control weight and the local control weight respectively, ΔX g is the update increment of the global optimization control parameter, and ΔX l is the update increment of the local optimization control parameter;

[0240] S53. Optimize the control weights α g ,α l using an adaptive weight adjustment strategy, and calculate the optimal control weights according to the real-time mixing uniformity U, the particle agglomeration rate C, and the energy consumption E:

[0241]

[0242] α l = 1 - α g

[0243] where w U ,w C ,w E are the adjustment weights of the mixing uniformity, the particle agglomeration rate, and the energy consumption respectively;

[0244] S54. Optimize the air flow pulse control strategy by combining the temperature change, and define the adjustment function of the air flow pulse control parameter with respect to the temperature:

[0245]

[0246] Among them, β T is the temperature adjustment coefficient, is the actual temperature of the current air flow pulse region, T * is the optimal temperature parameter calculated by local optimization;

[0247] When the temperature deviation satisfies the stable condition:

[0248]

[0249] Among them, ∈ T is the temperature adjustment threshold, then stop adjusting the temperature control parameter, otherwise continue to optimize;

[0250] S55. Adopt a closed-loop feedback control strategy to adjust the air flow pulse parameters in real time. After each optimization iteration period t, calculate the error correction term:

[0251]

[0252] Among them, γ1 is the feedback adjustment gain, is the ideal air flow pulse control parameter;

[0253] Finally, update the control parameter:

[0254]

[0255] S56. When the optimization converges, output the optimal air flow pulse control parameter

[0256]

[0257] Among them, is the optimal air flow pulse pressure finally optimized, is the optimal air flow pulse frequency finally optimized, is the optimal air flow velocity finally optimized, is the optimal air flow action angle finally optimized, is the optimal air flow disturbance frequency finally optimized, is the optimal air flow shear strength finally optimized, is the optimal temperature control parameter finally optimized.

[0258] A pharmaceutical air flow pulse mixing system based on an improved ant colony algorithm is used to execute a pharmaceutical air flow pulse mixing method based on an improved ant colony algorithm.

[0259] Example 1:

[0260] This embodiment takes the production process of dry powder inhalant preparations in a pharmaceutical enterprise as the application scenario. The enterprise mainly produces micro-nano particle drug preparations for the treatment of asthma and chronic obstructive pulmonary disease. The key technology of dry powder inhalant preparations lies in how to ensure the uniform distribution of micro-nano particles and avoid the problem of unstable drug efficacy caused by agglomeration or uneven mixing. In the actual production process, the enterprise uses the air flow pulse mixing technology to disperse and uniformly transport the micro-nano particles to ensure the stability of the particles in the inhalation device. However, in the traditional mixing method, there are problems such as serious particle agglomeration, poor mixing uniformity, and difficult optimization of air flow pulse parameters, resulting in unstable drug inhalation dose and affecting the actual treatment effect of patients.

[0261] The method of the present invention is tested on the production line of dry powder inhalant preparations in a pharmaceutical enterprise. The test sample is the mixed powder of salmeterol fluticasone micro-nano particles with a particle size between 1 - 5 μm, which is required to be uniformly distributed on the lactose carrier and achieve uniform mixing through the air flow pulse mixing method. In this embodiment, the method of the present invention is applied to this production line and compared with the traditional air flow pulse mixing method to verify the actual effect of the present invention in improving mixing uniformity, reducing the particle agglomeration rate, and optimizing energy consumption.

[0262] Experimental environment and equipment:

[0263] Experimental time: January 2024 - March 2024;

[0264] Experimental location: XX Pharmaceutical Company production line laboratory;

[0265] Mixing equipment: Industrial air flow pulse mixing system (frequency adjustable range: 0 - 200 Hz, air flow pressure adjustable range: 0 - 0.8 MPa);

[0266] Micro-nano particle sample: Salmeterol fluticasone powder, particle size distribution: 1 - 5 μm;

[0267] Mixing objective: Improve the mixing uniformity of micro-nano particles, reduce particle agglomeration, optimize air flow parameters, and improve production stability;

[0268] Before mixing, first collect the physical property parameters of pharmaceutical micro-nano particles, including particle size distribution, particle density, particle surface charge, and adhesion characteristics, and record the data set:

[0269]

[0270] Among them, the physical property data of 5000 sample particles are as follows: average particle size: δ = 2.3 ± 0.5 μm, particle density: ρ = 1.4 ± 0.1 g / cm3, particle surface charge: q = -23.5 ± 4.2 mV, particle adhesion characteristics: ψ = 0.65 ± 0.12;

[0271] Meanwhile, collect the initial control parameters of the airflow pulse, including airflow pressure, pulse frequency, pulse duration, and airflow velocity:

[0272]

[0273] Among them, the initial airflow pressure: P = 0.45 ± 0.05 MPa, the initial pulse frequency: f = 80 ± 10 Hz, the initial pulse duration: T = 5 ± 1 ms, and the initial airflow velocity: v = 2.5 ± 0.4 m / s;

[0274] Global optimization stage: Use the improved ant colony algorithm to optimize the global airflow pulse control parameters and calculate the best combination to improve the overall mixing uniformity. Finally, obtain:

[0275]

[0276] Local optimization stage: Use the chaotic fruit fly optimization algorithm to optimize the local transport characteristics of micro-nano particles, including shear strength, perturbation frequency, and action angle, to reduce particle agglomeration. Finally, obtain:

[0277]

[0278] Temperature control adjustment: Combine the influencing factors of temperature change and, through temperature self-adaptive optimization, keep the temperature stable during the mixing process at

[0279] The comparison of the mixing effects between the traditional method and the method of the present invention is as follows:

[0280]

[0281]

[0282] During the production process of 100 batches, the pharmaceutical mixing stability of the method of the present invention is far better than that of the traditional method. This embodiment shows that the method of the present invention effectively improves the mixing uniformity of particles, reduces the particle agglomeration rate, optimizes the energy consumption of airflow pulse parameters, and enhances the temperature control stability in the actual pharmaceutical production environment, meeting the high requirements of modern pharmaceutical processes for the uniform mixing of micro-nano particles.

[0283] The experimental results show:

[0284] 1. The mixing uniformity is improved by 16.7%, ensuring the stability of the drug inhalation dose.

[0285] 2. The particle agglomeration rate is reduced by 31.3%, reducing the decrease in inhalation efficiency caused by particle agglomeration.

[0286] 3. The energy consumption is reduced by 20%, the production energy efficiency is improved, and the equipment energy consumption is reduced.

[0287] 4. The temperature fluctuation is reduced by 65.6%, ensuring the stability of the mixing process and avoiding uneven distribution of particles caused by temperature effects.

[0288] The experimental results of the method of the present invention show that the optimization strategy not only improves the particle mixing quality, but also enhances the intelligence and stability of the production process, providing an efficient and reliable micro-nano particle mixing optimization solution for pharmaceutical enterprises.

[0289] The present invention combines the improved ant colony algorithm and the chaotic fruit fly optimization algorithm to achieve the dynamic coordination of global optimization and local optimization. In the global optimization stage, the improved ant colony algorithm optimizes the global control parameters of the air flow pulse through the adaptive pheromone initialization, particle dynamics heuristic factor, and dynamic evaporation coefficient adjustment strategy, improving the convergence speed and global search ability of the optimization algorithm. In the local optimization stage, the fruit fly optimization algorithm based on chaotic perturbation is introduced. Using the chaotic mapping perturbation mechanism and the local search adaptive step size control, the particle transport behavior is refined and optimized, enhancing the stability and accuracy of the local search. By combining the two optimization algorithms, the optimization path has strong global exploration ability in the initial stage, and at the same time, the search accuracy is improved through local refinement optimization in the later stage, avoiding the search falling into local optimum and improving the effectiveness of the air flow pulse parameter optimization.

[0290] The present invention introduces an adaptive temperature control mechanism in the optimization process, and adjusts the temperature influence factors in the air flow pulse mixing process in real time, solving the problem that the existing optimization methods ignore the influence of temperature changes on particle mixing behavior. In the mixing optimization process, the temperature influence factor is combined with the particle agglomeration rate for dynamic adjustment, so that the air flow pulse control parameters can adapt to different temperature changes, thereby improving the mixing stability of particles under different working conditions. The temperature-particle mixing coupling optimization strategy is adopted, and the temperature parameters are dynamically adjusted through a closed-loop feedback mechanism to make it meet the optimal mixing state, ensuring that the distribution of particles in the air flow is more uniform.

[0291] The present invention constructs a closed-loop control strategy based on adaptive feedback adjustment through the dynamic coordination of global optimization control parameters and local optimization control parameters, solving the problem that the existing optimization methods lack real-time adjustment ability and cannot dynamically adjust parameters according to the particle mixing state. An adaptive weight adjustment mechanism is introduced in the optimization process. According to the mixing uniformity, particle agglomeration rate, and energy consumption monitored in real time, the optimal global and local optimization control weights are calculated to realize the intelligent allocation of parameters, enabling the optimization algorithm to perform adaptive adjustment according to real-time changes.

[0292] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent replacements or changes, shall be covered by the protection scope of the present invention.

Claims

1. A pharmaceutical airflow pulse mixing method based on an improved ant colony algorithm, characterized in that: The steps include: S1. Collecting physical property parameters of pharmaceutical micro-nano particles and initial control parameters of airflow pulses to form a mixed standardized data set of pharmaceutical micro-nano particles and airflow pulses; S2. Establishing a pharmaceutical micro-nano particle airflow pulse mixing process optimization model based on the pharmaceutical micro-nano particle airflow pulse mixing standardized data set; S3. Use the improved ant colony algorithm to perform a global search on the objective function of the optimization model of the pharmaceutical micro-nano particle airflow pulse mixing process, optimize the global control parameters of the airflow pulse, and output the global optimization control parameters; S4. Based on the global optimization control parameter output, the chaotic fruit fly optimization algorithm is used to refine and optimize the local transport behavior of pharmaceutical micro-nano particles in the airflow to form a local optimization control parameter output; S5. According to the global optimization control parameter output and the local optimization control parameter output, the airflow pulse control parameter is dynamically adjusted in real time to form a closed-loop control strategy in which the global and local control parameters work together; S6. Apply the airflow pulse control parameters after the real-time dynamic adjustment to the pharmaceutical micro-nano particle airflow pulse mixing system to perform mixing.

2. The pharmaceutical airflow pulse mixing method based on improved ant colony algorithm according to claim 1 is characterized in that: The S1 comprises the following steps: S11. Collect the physical property parameters of pharmaceutical micro-nano particles and establish a data set D of the physical property parameters of pharmaceutical micro-nano particles p : Among them, D p Represents the physical property parameter data set of pharmaceutical micro-nanoparticles, N p is the number of pharmaceutical micro-nanoparticle samples collected, d p,i is the physical property data of the i-th pharmaceutical micro-nanoparticle, δ i represents the particle size of the i-th pharmaceutical micro-nano particle, ρ i represents the density of the i-th pharmaceutical micro-nanoparticle, q i represents the surface charge of the i-th pharmaceutical micro-nanoparticle, ψ i represents the adhesion characteristics of the i-th pharmaceutical micro-nanoparticle; S12. Collect the initial control parameters of the airflow pulse and establish the initial control parameter data set D of the airflow pulse f : Among them, D f represents the initial control parameter data set of the airflow pulse, N f is the number of airflow pulse parameter samples collected, d f,j is the initial control parameter of the jth group of airflow pulses, P j represents the pressure of the jth group of airflow pulses, f j represents the frequency of the jth group of airflow pulses, T j represents the duration of the jth group of airflow pulses, v j represents the flow rate of the jth group of airflow pulses; S13. The physical property parameter data set D of pharmaceutical micro-nano particles p and the initial control parameter data set D of the airflow pulse f The initial data set D of pharmaceutical micro-nano particle airflow pulse mixing is formed by association. m : D m ={(d p,i ,d f,j )∣d p,i ∈D p ,d f,j ∈D f ,R(d p,i ,d f,j )} Among them, D m represents the initial data set of pharmaceutical micro-nano particle airflow pulse mixing, R(d p,i ,d f,j ) is the correlation function between the physical properties of pharmaceutical micro-nano particles and the airflow pulse control parameters, which is constructed based on the influence of particle size, particle density, surface charge and adhesion characteristics on airflow pulse pressure, frequency, duration and flow rate; S14. Normalize the initial data set of pharmaceutical micro-nano particle airflow pulse mixing to generate a standardized data set of pharmaceutical micro-nano particle airflow pulse mixing And build the index.

3. The pharmaceutical airflow pulse mixing method based on improved ant colony algorithm according to claim 1, characterized in that: The S2 comprises the following steps: S21. Standardized data set based on pharmaceutical micro-nanoparticle airflow pulse mixing Establishment of an optimization model for the airflow pulse mixing process of pharmaceutical micro-nano particles o : M o =(X,Y,F); Where X is the input parameter vector of the pharmaceutical micro-nano particle airflow pulse mixing process, which is defined as: in, is the standardized particle size of the i-th pharmaceutical micro-nano particle, is the density of the i-th pharmaceutical micro-nanoparticle after standardization, is the surface charge of the ith pharmaceutical micro-nanoparticle after standardization, is the standardized adhesion property of the ith pharmaceutical micro-nanoparticle, P j * is the normalized pressure of the jth group of airflow pulses, f j * is the frequency of the jth group of airflow pulses after standardization, T j * is the duration of the jth group of airflow pulses after standardization, is the normalized flow rate of the jth group of airflow pulses; S22. Define the target output parameter vector Y of pharmaceutical micro-nano particle airflow pulse mixing: Y=(U,C,E); Among them, U represents the mixing uniformity index of pharmaceutical micro-nano particles, C represents the agglomeration rate index of pharmaceutical micro-nano particles, and E represents the energy consumption of the airflow pulse mixing process of pharmaceutical micro-nano particles; S23. Calculate the pharmaceutical micro-nanoparticle mixing uniformity index U: Where N is the number of sampling points in the mixed area, σ i is the standard deviation of particle concentration at the i-th sampling point, is the standard deviation of the average particle concentration in the entire mixing area; Analyze the particle concentration standard deviation σ by combining pharmaceutical micro-nano particle airflow pulse mixing standardized data set i Factors affecting particle size Affects the settling rate of particles in the airflow, thereby affecting local concentration fluctuations and particle density The drift behavior of particles in pulsed airflow and the surface charge of particles Affects the electrostatic interaction between particles, resulting in local agglomeration, which in turn affects the concentration standard deviation and particle adhesion characteristics The air flow pressure P affects the adhesion behavior of particles during the mixing process. j * The air flow pulse frequency f affects the transport intensity of particles. j * The periodic disturbance of the airflow on the particles causes the particles to be redistributed. The duration of the airflow pulse is T j * Affects the residence time of particles in the mixing area, thereby changing the local concentration fluctuation and air flow rate Affects the diffusion rate of particles; S24. Calculate the aggregation rate index C of pharmaceutical micro-nanoparticles: Where M is the number of divisions of the particle agglomeration region, n j represents the number of particles in the jth agglomeration area, N j Represents the theoretical number of uniformly distributed particles in the area, n j Surface charge of pharmaceutical micro-nanoparticles and adhesion properties Influence; S25. Calculate the energy consumption E of the pharmaceutical micro-nano particle airflow pulse mixing process: Among them, K is the number of airflow pulses, is the normalized pressure of the kth airflow pulse, is the normalized action time of the kth airflow pulse; the energy consumption E is affected by the airflow pulse frequency f j * and flow rate Influence; S26. Construct the optimization objective function F(X) of pharmaceutical micro-nano particle airflow pulse mixing process: F(X)=w1U-w2C-w3E Among them, w1, w2, w3 are optimization weight coefficients; S27. Based on the optimization objective function F(X) of the pharmaceutical micro-nano particle airflow pulse mixing process, the mixing behavior of the pharmaceutical micro-nano particles is simulated and calculated, an optimization solution framework is established, and the optimization model of the pharmaceutical micro-nano particle airflow pulse mixing process is updated.

4. The pharmaceutical airflow pulse mixing method based on improved ant colony algorithm according to claim 1 is characterized in that: The S3 comprises the following steps: S31. Based on the optimization objective function F(X) of the pharmaceutical micro-nano particle airflow pulse mixing process, the optimization search space Ω is defined: Among them, P j * is the normalized airflow pulse pressure, f j * is the standardized airflow pulse frequency, is the standardized air flow velocity; S32. Adopting the adaptive pheromone initialization mechanism based on the input parameter vector X, constructing the adaptive pheromone matrix τ based on the physical properties of pharmaceutical micro-nano particles and airflow pulse control parameters (0) : in, is the initial pheromone concentration at path (i, j), τ0 is the initial pheromone concentration baseline value, indicating the pheromone level before optimization, The surface charge of the standardized pharmaceutical micro-nano particles i affects the electrostatic interaction between particles, determines the tendency of particle agglomeration, and further affects the mixing uniformity U. is the average surface charge of pharmaceutical micro-nanoparticles, is the standard deviation of the surface charge of pharmaceutical micro-nanoparticles, is the mean value of the airflow pulse pressure, is the standard deviation of the airflow pulse pressure, is the mean velocity of the airflow pulse, is the standard deviation of the airflow pulse velocity; S33. Define the heuristic factor η at path (i,j) ij : in, The mean value of the sedimentation velocity, ξ ij Indicates the settling velocity of particles under the action of airflow pulse: in, is the particle density, which affects the floating properties of particles under the action of airflow pulses. The heuristic factor makes ants tend to choose the path that can improve the mixing uniformity U and reduce the particle agglomeration rate C when searching for the path; S34. Calculate the transition probability of ant k in each iteration Among them, α and β represent the weight parameters of pheromone intensity and heuristic factor, respectively, which make the search tend to optimize the direction of objective function F(X); S35. Combine the group behavior of micro-nano particles to dynamically update pheromones, so that the pheromone update is not only affected by the objective function, but also combined with the surface charge of micro-nano particles and flow rate Then optimize the search direction and define the pheromone update as: in, is the pheromone concentration of the ant at the path (i, j) in the t+1th iteration, indicating the preference of the ant on the path, ρ is the pheromone volatility coefficient, is the pheromone concentration at path (i, j) in the tth iteration, inheriting the pheromone level of the previous round, m is the number of ants participating in the pheromone update, The increment of pheromone released by the kth ant on the path (i, j): Among them, Q is the total pheromone release coefficient, is the fitness value of the kth ant at the path (i, j), indicating the contribution of the path to the optimization goal. The larger the fitness value, the better the path and the stronger the pheromone update. S36. Using dynamic volatility coefficient to optimize search efficiency: Among them, ρ min and ρ max are the set minimum and maximum volatility coefficients, λ1 is the adjustment rate parameter, and t is the current iteration number; S37. Output global optimization control parameters:

5. The pharmaceutical airflow pulse mixing method based on improved ant colony algorithm according to claim 1, characterized in that: The S4 comprises the following steps: S41. Construct a local optimization search space based on the global optimization control parameters and define the optimization search space Ω of the local hybrid control parameters ′ : in, Indicates the local optimized airflow pulse action angle, represents the local airflow disturbance frequency, represents the local air flow shear intensity, represents the temperature of the local particle mixing area; S42. Use the chaotic fruit fly optimization algorithm based on global optimization parameters to initialize the local optimization population, and set the fruit fly population size to N f , and optimize the search space Ω locally ′ Randomly initialize the internal data so that the initial individuals satisfy: in, is the individual position of the i-th fruit fly. By introducing the global optimization control parameters (P, f, v), the initial position of the fruit fly population is perturbed and adjusted so that its search direction is consistent with the global optimization result: Among them, ∈1 is the local search perturbation factor, which makes the initial search of the fruit fly closer to the optimal solution; S43. Calculate the position of fruit fly individual i based on the target output parameter Local airflow mixing performance indicators at: Among them, T opt is the temperature regulation fitness, w1,w2,w3,w4 are weight coefficients; S44. Use the fruit fly olfactory search mechanism for local optimization, and each fruit fly updates its current position according to its own airflow disturbance characteristics: Among them, λ f is the local optimization step size factor, is the current local optimal solution; S45. Adopt the local search strategy based on chaotic perturbation. In each iteration, introduce the chaotic map to perform local perturbation: Among them, μ(t) is the dynamic chaotic disturbance factor, is the current average position of all fruit fly individuals, ξ is a random variable uniformly distributed between [0,1]; S46. Optimize particle mixing behavior in combination with temperature changes and define the factors affecting particle mixing: Among them, T eff is the actual acting temperature of the particles, is the temperature setting value of the current fruit fly individual, C is the particle aggregation rate, the higher the aggregation rate, the greater the effect of temperature on mixing; When the temperature stability condition is met, the current optimization state is acceptable, otherwise the temperature adjustment optimization is performed: |T eff -T|<∈ T ; Among them, ∈ T is the temperature stability threshold; S47. When the termination condition is met, the optimal local hybrid control parameters are output: in, To optimize the optimal local airflow angle, To optimize the optimal local airflow disturbance frequency, To optimize the optimal local airflow shear intensity, The optimal local temperature control parameters obtained by optimization.

6. The pharmaceutical airflow pulse mixing method based on improved ant colony algorithm according to claim 1, characterized in that: The S5 comprises the following steps: S51. Based on the global optimization control parameter output and the local optimization control parameter output, the airflow pulse dynamic control model is constructed, and the optimization variable vector X of the airflow pulse control parameter is defined. c : Constructing the airflow pulse dynamic control model M c : M c =(X c ,F c ) Among them, F c It is a real-time dynamic control strategy function; S52. Adopt global and local coordinated control strategy to realize adaptive adjustment of airflow pulse and define dynamic adjustment function of global and local coordinated control: in, is the airflow pulse control parameter of the t+1th iteration, α g and α l are the global control weight and the local control weight, ΔX g is the update increment of the global optimization control parameter, ΔX l The update increment of the local optimization control parameters; S53. Adopting adaptive weight adjustment strategy to optimize control weight α B ,α l , the optimal control weight is calculated according to the real-time mixing uniformity U, particle agglomeration rate C and energy consumption E: α l ＝1-α g Among them, w U ,w C ,w E are the adjustment weights of mixing uniformity, particle agglomeration rate and energy consumption, respectively; S54. Optimize the airflow pulse control strategy in combination with temperature changes, and define the adjustment function of the airflow pulse control parameters with temperature: Among them, β T is the temperature regulation coefficient, is the actual temperature of the current airflow pulse area, T * The optimal temperature parameters calculated for local optimization; When the temperature deviation meets the stability condition: Among them, ∈ T If it is the temperature adjustment threshold, stop adjusting the temperature control parameters, otherwise continue to optimize; S55. Use a closed-loop feedback control strategy to adjust the airflow pulse parameters in real time. After each optimization iteration cycle t, calculate the error correction term: Where γ1 is the feedback adjustment gain, is the ideal airflow pulse control parameter; Finally update the control parameters: S56. When the optimization converges, output the optimal airflow pulse control parameters in, To obtain the optimal airflow pulse pressure, The optimal airflow pulse frequency is finally optimized. To obtain the optimal airflow velocity, The optimal airflow angle obtained by the final optimization is: The optimal airflow disturbance frequency obtained by the final optimization is: To finally optimize the optimal airflow shear strength, The optimal temperature control parameters are finally obtained by optimization.

7. A pharmaceutical airflow pulse mixing system based on an improved ant colony algorithm, characterized in that: Used to execute a pharmaceutical airflow pulse mixing method based on an improved ant colony algorithm as described in any one of claims 1-6.

Images