Process optimization method for realizing extraction of traditional Chinese medicine components, extraction method, device and medium

Abstract

The invention discloses a process optimization method for realizing traditional Chinese medicine component extraction, a traditional Chinese medicine component extraction method, a device and a medium. A continuous process parameter space is constructed according to characteristic data of traditional Chinese medicine raw materials; establishing a process trajectory function based on the continuous process parameter space; performing simulation extraction on the ant colony individuals according to the trajectory control node parameters corresponding to the ant colony individuals, and generating multi-target fitness values corresponding to the ant colony individuals; and performing iterative search on the trajectory control node parameters in the continuous process parameter space according to the multi-target fitness values of the ant colony individuals, and outputting a target trajectory control node parameter set. Therefore, by constructing the continuous process parameter space and utilizing the ant colony algorithm to carry out iterative search, trajectory control node parameters are dynamically optimized, and the problem that a traditional process cannot adapt to raw material variation is effectively solved; the method has the advantages of dynamically adapting to the characteristic change of the traditional Chinese medicine raw materials, realizing multi-target collaborative optimization and improving the extraction efficiency and the product quality stability.

Description

Technical Field

[0001] This invention relates to the field of traditional Chinese medicine pharmaceutical technology, specifically to a process optimization method, a method for extracting traditional Chinese medicine components, an apparatus, and a computer-readable storage medium for this purpose. Background Technology

[0002] Traditional Chinese medicine (TCM), as an important part of the traditional medicine system, faces many technical bottlenecks in modern production processes. Traditional TCM extraction processes generally use preset fixed process parameters, such as uniform temperature control, solvent concentration ratio, and stirring speed settings. This "one-size-fits-all" approach fails to fully consider the inherent natural variation characteristics of TCM materials.

[0003] Due to significant differences in planting regions, growth cycles, harvesting seasons, and storage conditions, different batches of raw materials exhibit high inconsistencies in moisture content, cellulose structure distribution, tissue density, and initial concentrations of target active ingredients. Extraction methods with fixed parameters cannot dynamically adapt to changes in raw material characteristics, leading to drastic fluctuations in the extraction efficiency of target components and poor batch-to-batch product quality stability, making it difficult to meet the stringent requirements of modern pharmaceutical production for purity, yield, and safety. Furthermore, existing process optimization methods primarily rely on operators' subjective experience or offline experimental design methods, such as statistical response surface methodology or orthogonal experimental design. These methods require repeated and extensive physical experiments to determine parameter combinations, which is not only time-consuming and resource-intensive but also fails to allow for real-time monitoring and timely adjustments to the dynamic fluctuations in raw material characteristics during production.

[0004] Traditional automated control systems, while capable of executing preset programs, lack intelligent decision-making mechanisms for the complex extraction systems of traditional Chinese medicine (TCM) and the multi-objective coordination (such as maximizing the yield of target components, minimizing impurity dissolution, and reducing energy consumption) and multiple constraints (such as equipment capacity limits and raw material characteristic limitations). In actual production, the pursuit of high yield often forces the sacrifice of purity control or increases in energy consumption, leading to problems such as increased impurity content, higher subsequent purification costs, and energy waste. This makes it difficult to achieve a dynamic balance between multiple objectives in process parameters. Furthermore, the adjustment of process parameters relies excessively on manual intervention and experience accumulation. When faced with sudden changes in raw materials or equipment malfunctions, the system response is sluggish, failing to achieve adaptive adjustment and refined control of the production process. This severely restricts the intelligent upgrading and standardization of TCM production. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing technology and solve the existing technical problems, the present invention provides a process optimization method, extraction method, process optimization device, extraction device and computer storage medium for realizing the extraction of Chinese medicine components. It has the advantages of dynamically adapting to changes in the characteristics of Chinese medicine raw materials, realizing multi-objective synergistic optimization, improving extraction efficiency and product quality stability.

[0006] In a first aspect, embodiments of the present invention provide a process optimization method for extracting components from traditional Chinese medicine. The method includes: constructing a continuous process parameter space based on characteristic data of the traditional Chinese medicine raw materials; establishing a process trajectory function based on the continuous process parameter space; simulating extraction of an ant colony individual based on trajectory control node parameters corresponding to the ant colony individual, generating a multi-objective fitness value corresponding to the ant colony individual; iteratively searching the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of the ant colony individual, and outputting a set of target trajectory control node parameters.

[0007] In one technical solution, a continuous process parameter space is constructed based on the characteristic data of Chinese herbal raw materials, including: collecting characteristic data of Chinese herbal raw materials; determining the solid-liquid ratio of the Chinese herbal raw materials based on the collected characteristic data; reading the capacity boundary parameters of the extraction equipment to form a set of equipment capacity boundaries; and constructing a continuous process parameter space by combining the characteristic vector of the Chinese herbal raw materials and the obtained solid-liquid ratio under the constraints of the set of equipment capacity boundaries.

[0008] In one technical solution, the process of simulating and extracting the trajectory control node parameters corresponding to the individual ant colony members to generate a multi-objective fitness value for the individual ant colony members includes: reading the trajectory control node parameters corresponding to the individual ant colony members to form a set of trajectory control node parameters; generating the process trajectory of the individual ant colony members within the extraction time range based on the formed set of trajectory control node parameters; and using the generated process trajectory of the individual ant colony members as input boundary conditions to simulate and extract the individual ant colony members in a model simulation environment to generate a multi-objective fitness value for the individual ant colony members.

[0009] In one technical solution, the trajectory control node parameters are iteratively searched in the continuous process parameter space based on the multi-objective fitness values ​​of the ant colony individuals to output a target trajectory control node parameter set. This includes: updating the pheromone matrix based on the multi-objective fitness values ​​of the ant colony individuals to obtain an updated pheromone matrix; and using the updated pheromone matrix, iteratively searching the trajectory control node parameters in the continuous process parameter space to output a target trajectory control node parameter set that satisfies the convergence condition.

[0010] In one technical solution, updating the pheromone matrix based on the multi-objective fitness values ​​of individual ant colonies to obtain the updated pheromone matrix includes: establishing a mapping relationship between the pheromone matrix and control node parameters based on the multi-objective fitness values ​​of individual ant colonies; determining the pheromone deposition amount for the corresponding control node parameters based on the established mapping relationship between the pheromone matrix and control node parameters; performing pheromone volatilization processing on the pheromone matrix; and updating the pheromone matrix after pheromone volatilization processing based on the determined pheromone deposition amount to obtain the updated pheromone matrix.

[0011] In one technical solution, the step of using the updated pheromone matrix to iteratively search the trajectory control node parameters in the continuous process parameter space and outputting a target trajectory control node parameter set that satisfies the convergence condition includes: establishing a transition selection rule based on the updated pheromone matrix; iteratively searching the trajectory control node parameters in the continuous process parameter space according to the established transition selection rule; determining the convergence condition during the iterative search process; and outputting the target trajectory control node parameter set when the convergence condition is satisfied.

[0012] Secondly, embodiments of the present invention also provide a method for extracting components of traditional Chinese medicine, the method comprising: converting a set of trajectory control node parameters into a set of control instructions for extracting components of traditional Chinese medicine, wherein the set of trajectory control node parameters is obtained by outputting a process optimization method for realizing the extraction of components of traditional Chinese medicine as described in any of the above technical solutions; collecting extraction conditions corresponding to the raw materials of traditional Chinese medicine for component extraction; and executing the set of control instructions according to the extraction conditions to extract the target component.

[0013] Thirdly, embodiments of the present invention also provide a process optimization device for extracting components from traditional Chinese medicine. The device includes: a space construction module for constructing a continuous process parameter space based on the characteristic data of the traditional Chinese medicine raw materials; a function establishment module for establishing a process trajectory function based on the continuous process parameter space; a simulated extraction module for simulating extraction of an ant colony individual based on the trajectory control node parameters corresponding to the ant colony individual, generating a multi-objective fitness value corresponding to the ant colony individual; and an iterative search module for iteratively searching the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of the ant colony individual, outputting a set of target trajectory control node parameters.

[0014] Fourthly, embodiments of the present invention also provide a traditional Chinese medicine component extraction device, the device comprising: a conversion module, configured to convert a set of trajectory control node parameters into a set of control instructions for traditional Chinese medicine component extraction, the set of trajectory control node parameters being output by the process optimization device for realizing traditional Chinese medicine component extraction as described above; a data acquisition module, configured to acquire the extraction conditions corresponding to the traditional Chinese medicine raw materials for component extraction; and a component extraction module, configured to execute the set of control instructions according to the extraction conditions to extract the target component.

[0015] Fifthly, embodiments of the present invention also provide a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are invoked or executed by a processor, the computer-executable instructions cause the processor to implement the above-mentioned process optimization method or method for extracting traditional Chinese medicine components.

[0016] This invention provides a process optimization method, extraction method, apparatus, and computer-readable storage medium for extracting components from traditional Chinese medicine (TCM). The method involves constructing a continuous process parameter space based on the characteristic data of the TCM raw materials; establishing a process trajectory function based on this continuous process parameter space; simulating extraction on individual ant colonies according to their corresponding trajectory control node parameters to generate multi-objective fitness values ​​for each ant colony individual; and iteratively searching the trajectory control node parameters within the continuous process parameter space based on these multi-objective fitness values ​​to output a set of target trajectory control node parameters. Thus, this invention, by constructing a continuous process parameter space and utilizing an ant colony algorithm for iterative search, dynamically optimizes the trajectory control node parameters, effectively solving the problem of traditional processes being unable to adapt to raw material variations. It offers advantages such as dynamically adapting to changes in the characteristics of TCM raw materials, achieving multi-objective collaborative optimization, and improving extraction efficiency and product quality stability. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a process optimization method for extracting components from traditional Chinese medicine, provided in an embodiment of the present invention. Figure 2 This is a schematic flowchart of the extraction method for traditional Chinese medicine components provided in an embodiment of the present invention; Figure 3 A schematic diagram of the composition structure of the process optimization device for extracting traditional Chinese medicine components provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the composition and structure of the traditional Chinese medicine component extraction device provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without inventive effort are within the scope of protection of the present application.

[0019] Traditional Chinese medicine extraction methods often employ fixed process parameters, ignoring batch-to-batch variations in medicinal materials, leading to unstable extraction efficiency and product quality. Existing optimization methods largely rely on experience or offline experiments, making it difficult to dynamically respond to changes in raw materials. Furthermore, traditional automated control systems lack the intelligent decision-making capabilities to handle the multi-objective and multi-constraint nature of complex Chinese medicine systems. This makes it difficult to achieve multi-objective balance in the process, parameter adjustments are highly dependent on manual intervention, and the production process response is sluggish, hindering intelligent and precise control.

[0020] To address the aforementioned shortcomings, embodiments of the present invention provide a process optimization method for extracting components from traditional Chinese medicine, such as... Figure 1 As shown, the process includes the following steps: S101, constructing a continuous process parameter space based on the characteristic data of Chinese medicinal materials; S102, establishing a process trajectory function based on the continuous process parameter space; S103, performing simulation extraction on the ant colony individuals based on the trajectory control node parameters corresponding to the ant colony individuals, generating multi-objective fitness values ​​corresponding to the ant colony individuals; S104, performing iterative search on the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of the ant colony individuals, and outputting a set of target trajectory control node parameters.

[0021] In step S101 above, specifically, key process parameters during the extraction process can be preset, such as extraction temperature, solvent concentration, stirring speed, and extraction time, and a fixed operating range can be set for each parameter. These operating ranges can be initially determined based on historical experience data or industry standards. For example, for a certain type of Chinese medicinal material, its extraction temperature range may be set to 60℃ to 80℃, and the solvent concentration range to 50% to 70%. The characteristic data of the Chinese medicinal raw materials can be simply identified by batch number or preset classification label, and then the corresponding preset parameter space can be selected according to these labels. The characteristic data of the Chinese medicinal raw materials refers to the quantitative information used to describe the physical and chemical properties and target component content of the Chinese medicinal materials, such as the water content, cellulose content, tissue density, and initial content of target effective components of the Chinese medicinal materials. These data are used to reflect the inherent differences between different batches of Chinese medicinal materials. The continuous process parameter space refers to a multidimensional continuous region composed of various adjustable process parameters (such as temperature, solvent concentration, stirring speed, extraction time, etc.) during the extraction process. Any point in this space represents a set of feasible process parameter combinations used to guide the extraction process.

[0022] In step S102 above, the process trajectory function can be defined as a series of discrete process parameter points, which are linearly interpolated and connected along the extraction time axis to form a simple process path. For example, the extraction process can be set to start at 60°C, rise to 70°C halfway through extraction, and remain at 70°C at the end, with this temperature change described by a piecewise linear function. Solvent concentration or stirring speed can also be defined using similar piecewise linear or step functions. The process trajectory function refers to a mathematical function or curve that describes the dynamic changes of extraction process parameters over time in a continuous process parameter space. It defines the parameter change path from the start to the end of extraction to achieve refined control of the extraction process.

[0023] In step S103 above, each ant colony individual is assigned a set of trajectory control node parameters, which define the process trajectory it represents. For example, an ant colony individual might correspond to a simple trajectory of "starting temperature 60℃, ending temperature 75℃, solvent concentration 60%". Subsequently, a simplified empirical model or lookup table method is used to simulate the process trajectory to predict its potential extraction effects, such as the yield of the target component and the impurity content. These prediction results are combined into a multi-objective fitness value, for example, by a simple weighted summation method, to evaluate the merits of the process trajectory.

[0024] Here, the term "ant colony individual" refers to a virtual search agent that simulates ant behavior in the ant colony algorithm. Each ant colony individual represents a potential process optimization scheme and explores the optimal solution by moving within the parameter space. The "trajectory control node parameters" refer to the set of parameters used to define the key points of the process trajectory function. These node parameters determine the shape and path of the process trajectory, and adjusting these node parameters can change the dynamic behavior of the extraction process. "Simulated extraction" refers to the process of predicting the results of the extraction process (such as yield, impurity content, energy consumption, etc.) based on given process parameters and the characteristic data of the Chinese herbal medicine raw materials in a computer model or simulation environment, without the need for actual physical experiments. The "multi-objective fitness value" refers to a quantitative evaluation index calculated during the simulated extraction process for multiple optimization objectives (such as maximizing yield, minimizing impurities, minimizing energy consumption, etc.), used to measure the comprehensive performance of the process scheme represented by each ant colony individual.

[0025] In step S104 above, the iterative search process can employ a random walk or a simple selection-based strategy. For example, in each iteration, new ant colony individuals are randomly generated, and their fitness values ​​are simulated and extracted. Then, ant colony individuals with higher fitness values ​​are selected as the starting point for the next round of search, while individuals with lower fitness values ​​are discarded. This process is repeated until a preset number of iterations is reached or the fitness value no longer significantly improves, ultimately outputting the current optimal set of trajectory control node parameters.

[0026] Here, the iterative search refers to an optimization process that gradually approximates or finds the optimal solution by repeatedly executing a series of computational steps. In each iteration, the search direction or parameters are adjusted based on the results of the previous iteration until the preset convergence conditions are met. The target trajectory control node parameter set refers to the final combination of trajectory control node parameters determined after iterative search that can achieve the best or satisfactory multi-target extraction effect. This set can be directly used to guide the actual extraction process of traditional Chinese medicine components.

[0027] This invention provides a method for optimizing the extraction process of traditional Chinese medicine components. By constructing a process parameter space based on the characteristics of the raw materials and establishing a process trajectory function, the extraction process can be dynamically described. Through simulated extraction and multi-objective fitness value evaluation, multiple objectives such as yield and impurities can be comprehensively considered. Furthermore, by iteratively searching and optimizing the trajectory control node parameters, the efficiency and quality instability caused by batch differences in raw materials in traditional Chinese medicine extraction can be effectively solved, thereby reducing reliance on manual experience and improving the optimization capability of the process and the level of production precision.

[0028] In the above embodiments of the present invention, a continuous process parameter space is proposed to optimize the extraction process of traditional Chinese medicine. However, in its implementation, due to the lack of systematic collection of characteristic data of traditional Chinese medicine raw materials, dynamic determination of solid-liquid ratio, and effective integration of extraction equipment capacity boundaries, the constructed space may not accurately reflect the batch differences of raw materials and the actual constraints of equipment, which leads to a decrease in the reliability of subsequent process trajectory function establishment and optimization search, affecting the overall extraction efficiency and stability.

[0029] Based on this, in one embodiment of the present invention, the present invention further proposes a method for constructing a continuous process parameter space based on the characteristic data of traditional Chinese medicine raw materials during the implementation of step S101. The method includes: collecting characteristic data of traditional Chinese medicine raw materials; determining the solid-liquid ratio of the traditional Chinese medicine raw materials based on the collected characteristic data of the traditional Chinese medicine raw materials; reading the capacity boundary parameters of the extraction equipment to form a set of equipment capacity boundaries; and constructing a continuous process parameter space by combining the characteristic vector of the traditional Chinese medicine raw materials and the obtained solid-liquid ratio under the constraints of the set of equipment capacity boundaries.

[0030] Specifically, collecting characteristic data on Chinese medicinal materials aims to obtain their actual physical and chemical properties, ensuring that the construction of subsequent process parameter spaces can fully reflect batch-to-batch differences. For example, various analytical methods such as near-infrared spectroscopy (NIR), high-performance liquid chromatography (HPLC), moisture content determination, density measurement, and particle size distribution analysis can be used to obtain key characteristic data such as the content of effective components, impurities, water content, cellulose content, and tissue density of Chinese medicinal materials. Furthermore, image recognition technology can be used to collect macroscopic characteristics such as the morphology, color, and integrity of Chinese medicinal materials.

[0031] Here, determining the solid-liquid ratio of the collected Chinese herbal medicine raw materials based on their characteristic data is to achieve dynamic adjustment of the solid-liquid ratio, matching it to the characteristics of the current batch of raw materials, thereby optimizing extraction efficiency and solvent utilization. For example, based on a pre-established mathematical model or empirical formula, the collected characteristic data of the raw materials (such as water absorption rate, solubility of active ingredients, density, etc.) can be used as input to calculate the solid-liquid ratio that achieves the best extraction effect or highest yield under the characteristics of the raw materials. Alternatively, machine learning algorithms can be used to predict the optimal solid-liquid ratio for the current raw materials by training on the relationship between different raw material characteristics and their corresponding optimal solid-liquid ratios in historical data.

[0032] Reading the capacity boundary parameters of the extraction equipment to form a set of equipment capacity boundaries is to take into account the actual operating limitations of the extraction equipment and ensure that the constructed process parameter space is physically feasible. For example, parameters such as the equipment's maximum / minimum operating temperature, maximum / minimum operating pressure, stirring speed range, solvent flow rate range, and maximum throughput can be read from the equipment's control system or configuration database. These parameters constitute the operating boundaries of the equipment under safe and efficient operating conditions.

[0033] Furthermore, under the constraints of the equipment capability boundary set, and combining the feature vectors of the Chinese herbal raw materials with the obtained solid-liquid ratio, a continuous process parameter space is constructed. This aims to create an effective optimization region that can adapt to changes in raw materials while also meeting equipment limitations. For example, a multidimensional hypercube can be defined, where each dimension represents a process parameter (such as temperature, pressure, extraction time, solvent concentration, etc.). The upper and lower limits of these dimensions are first determined by the equipment capability boundary set, and then further adjusted and refined based on the feature vectors of the Chinese herbal raw materials (such as heat sensitivity and solubility characteristics) and the determined solid-liquid ratio. For example, if the raw material is sensitive to high temperatures, the upper temperature limit may be further reduced; if the adjustment of the solid-liquid ratio affects the effective contact area of ​​the solvent, the range of extraction time or stirring speed may be adjusted accordingly. In this way, the constructed continuous process parameter space is a dynamic, adaptive, and physically feasible search region.

[0034] Through the above technical solution, this invention systematically collects characteristic data of traditional Chinese medicine raw materials and dynamically determines the solid-liquid ratio based on this data. Simultaneously, it combines this data with the capacity boundary parameters of the extraction equipment to construct a continuous process parameter space that accurately reflects batch-to-batch differences in raw materials and the actual constraints of the equipment. This construction method avoids the optimization result deviations caused by inaccurate parameter spaces in traditional methods, ensuring the reliability of subsequent process trajectory function establishment and ant colony optimization search. Specifically, by collecting characteristic data of traditional Chinese medicine raw materials, the parameter space can be adaptively adjusted for the characteristics of different batches of raw materials. For example, for raw materials with high water content, the solid-liquid ratio and extraction time range can be adjusted accordingly. At the same time, the capacity boundary parameters of the extraction equipment are taken into consideration, ensuring that any process point in the constructed parameter space is achievable by the equipment, avoiding the generation of infeasible optimization schemes. This precise and dynamic parameter space construction method provides a solid foundation for subsequent process optimization based on ant colony algorithms, significantly improving the efficiency of traditional Chinese medicine component extraction and the stability of product quality, while effectively reducing the risks and resource consumption during the optimization process.

[0035] In the above embodiments of the present invention, it is also proposed to simulate and extract ant colony individuals based on the trajectory control node parameters corresponding to the ant colony individuals to generate multi-objective fitness values, which are used to evaluate the performance of ant colony individuals and support iterative search. However, in its implementation, there are technical defects in how to efficiently and accurately generate process trajectories based on trajectory control node parameters and perform simulation extraction to ensure the authenticity of fitness values ​​and the comprehensiveness of multi-objective evaluation. Specifically, the lack of systematic parameter processing leads to inaccurate trajectory generation, the unrealistic simulation environment affects the reliability of evaluation, and it is difficult to dynamically respond to multi-objective optimization requirements.

[0036] Based on this, in one embodiment of the present invention, during the implementation of step S103, the present invention further proposes to simulate and extract the ant colony individual based on the trajectory control node parameters corresponding to the ant colony individual, and generate a multi-objective fitness value corresponding to the ant colony individual. This process includes: reading the trajectory control node parameters corresponding to the ant colony individual to form a set of trajectory control node parameters; generating the process trajectory of the ant colony individual within the extraction time range based on the formed set of trajectory control node parameters; using the generated process trajectory of the ant colony individual as input boundary conditions, simulating and extracting the ant colony individual in a model simulation environment, and generating a multi-objective fitness value corresponding to the ant colony individual.

[0037] Specifically, the goal is to systematically acquire and organize the discrete control points that define the potential process trajectories of individual ants by reading the trajectory control node parameters corresponding to each individual ant. This ensures that all parameters required for trajectory generation are available in a structured format, avoiding omissions and facilitating subsequent processing. For example, the trajectory control node parameters carried by each individual ant can be read from the current iteration state of the ant colony optimization algorithm through a predefined data interface or file format (such as JSON, XML, or CSV). These parameters typically include the set values ​​of key process variables such as time points, temperature, pressure, and solvent flow rates at specific time points. After reading, these discrete parameter points are organized into an ordered list or array, thus forming a set of trajectory control node parameters. Alternatively, in the software implementation of the ant colony optimization algorithm, the data structure of individual ants is usually stored directly in memory. In this case, the data fields of the current individual ant can be accessed directly through pointers or object references to extract the trajectory control node parameters contained therein. These parameters can be encapsulated in a custom data structure, thus naturally forming a set.

[0038] Based on this, the process trajectories of individual ant colonies are generated within the extraction time range according to the formed set of trajectory control node parameters. This aims to transform the discrete set of control node parameters into a continuous, time-dependent process trajectory, which can be used as simulation input. This step ensures a smooth and accurate representation of process variables throughout the entire extraction duration by interpolating or approximating between defined control nodes. For example, various interpolation methods, such as linear interpolation, spline interpolation (e.g., cubic spline interpolation), or polynomial interpolation, can be used to generate a continuous process trajectory within the entire extraction time range based on discrete points in the trajectory control node parameter set. For the temperature parameter, if it is set to T1 at time t1 and T2 at time t2, the temperature value at time points between t1 and t2 can be calculated using an interpolation function. Alternatively, curve fitting techniques, such as least squares, can be used to fit the data points in the trajectory control node parameter set to obtain a continuous function that describes the process parameter's change over time. Evaluating this function within the extraction time range generates a smooth process trajectory.

[0039] Furthermore, the process trajectories of the generated ant colony individuals are used as input boundary conditions. The extraction process is simulated on these individuals in a model simulation environment, generating multi-objective fitness values ​​for each individual. This aims to evaluate the performance of the generated process trajectories under simulated conditions, thereby obtaining a quantitative measure reflecting the quality of the extraction process (multi-objective fitness value). This step is crucial for providing feedback to the optimization algorithm. For example, mathematical models based on the physicochemical principles of traditional Chinese medicine extraction processes (e.g., mass transfer kinetics models, thermodynamic models, fluid dynamics models, etc.) can be used as the simulation environment. The generated process trajectories (such as changes in temperature, pressure, and solvent flow rate over time) are used as input boundary conditions for these models. By numerically solving these model equations, key indicators such as the dissolution of target components, the precipitation of impurities, and energy consumption during the extraction process are simulated, thereby calculating multi-objective fitness values ​​(e.g., target component yield, impurity content, energy consumption, etc.). Alternatively, data-driven models based on historical experimental data or machine learning methods can be used as the simulation environment. For example, neural networks, support vector machines, or regression models can be used to learn the mapping relationship between process parameters and extraction results. The generated process trajectories are input into these data-driven models to predict the extraction effects that may be achieved under these process conditions, and multi-objective fitness values ​​are generated accordingly.

[0040] Through the above technical solution, this embodiment of the invention ensures the completeness and structure of parameter acquisition by reading trajectory control node parameters and forming a set, avoiding incomplete input problems caused by parameter omissions or scattered processing, thus providing a reliable foundation for subsequent steps. Simultaneously, based on the formed trajectory control node parameter set, a process trajectory is dynamically constructed within the extraction time range, considering time constraints to ensure the feasibility and accuracy of the trajectory, overcoming the mismatch risk caused by arbitrarily generated trajectories. Finally, the generated process trajectory is used as input boundary conditions for simulation extraction in a model simulation environment to generate multi-objective fitness values. The simulation environment simulates the real extraction process, using the trajectory as boundary conditions to ensure the realism of the simulation and the comprehensiveness of multi-objective evaluation, effectively improving the accuracy of fitness values ​​and the reliability of optimization decisions. Overall, this technical solution, through parameter settling, dynamic trajectory generation, and simulation environment application, collaboratively solves the uncertainty problem in the evaluation process, supports efficient multi-objective optimization, and enables the ant colony optimization algorithm to more accurately evaluate the performance of each individual, thereby conducting more effective iterative searches in the continuous process parameter space and ultimately outputting a better target trajectory control node parameter set.

[0041] In the above embodiments of this application, it is proposed to iteratively search the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of individual ant colonies in order to optimize the parameters and output the target set. However, in this process, there are challenges in how to efficiently update the pheromone matrix to dynamically reflect the quality of the search path and ensure that the search process can intelligently converge to the target parameter set that meets the predetermined conditions, so as to avoid low search efficiency or getting trapped in local optima.

[0042] Based on this, in one embodiment of the present invention, during the implementation of step S104, the present invention further proposes to iteratively search the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of individual ant colonies, and output a set of target trajectory control node parameters, including: updating the pheromone matrix based on the multi-objective fitness values ​​of individual ant colonies to obtain an updated pheromone matrix; and using the updated pheromone matrix to iteratively search the trajectory control node parameters in the continuous process parameter space, and output a set of target trajectory control node parameters that meets the convergence condition.

[0043] Specifically, regarding updating the pheromone matrix based on the multi-objective fitness values ​​of individual ants, this step aims to dynamically adjust the distribution of the pheromone matrix according to the multi-objective fitness values ​​obtained by individual ants during the simulated extraction process. The pheromone matrix is ​​a key mechanism in the ant colony algorithm used to store and transmit search experience; its value reflects the superiority or inferiority of different paths or parameter combinations. By updating the pheromone matrix, the algorithm can strengthen paths that lead to better multi-objective fitness values ​​and weaken poorly performing paths, thereby guiding subsequent search directions. One implementation is to calculate the amount of pheromone deposited on the traversed paths based on the fitness values ​​obtained by all ants or the best-performing ant after each iteration. Simultaneously, a certain proportion of all pheromones in the pheromone matrix undergoes evaporation to simulate the natural dissipation of pheromones, avoiding premature convergence to local optima and promoting exploration of new regions. Another implementation is to employ a ranking mechanism, allowing only the top-ranked ants to deposit pheromones, with higher-ranked individuals depositing larger amounts of pheromones. This method can more effectively utilize information from high-quality solutions, accelerating the algorithm's convergence to the global optimum. Furthermore, pheromone updates can be adaptive, for example, dynamically adjusting the pheromone evaporation rate and deposition amount based on the search's convergence speed or the diversity of solutions.

[0044] Furthermore, regarding the iterative search of the trajectory control node parameters in the continuous process parameter space using the updated pheromone matrix, outputting a set of target trajectory control node parameters that satisfy the convergence condition, this step utilizes the updated pheromone matrix to guide individual ants in a new round of trajectory control node parameter selection and exploration in the continuous process parameter space. Iterative search is an iterative process aimed at gradually optimizing the trajectory control node parameters to achieve better multi-objective fitness. Guided by the pheromone matrix, individual ants tend to select parameter regions that have historically proven effective, thereby improving search efficiency. One implementation is that when selecting the next trajectory control node parameter, individual ants calculate the selection probability based on the pheromone concentration of the corresponding parameter region in the pheromone matrix and heuristic information (e.g., distance to the current optimal solution). Parameter regions with higher pheromone concentrations or more favorable heuristic information are more likely to be selected. Through this probabilistic selection mechanism, the ant colony can effectively explore the parameter space. Another implementation is that, in each iteration, in addition to probabilistic selection based on pheromones, a local search strategy can also be introduced. For example, after an individual ant colony determines a set of trajectory control node parameters, it can be subjected to small-scale local perturbations or optimizations to further improve its multi-objective fitness. Iterative search will continue until preset convergence conditions are met, such as reaching the maximum number of iterations, the optimal solution not significantly improving over multiple generations, or the solutions generated by individual ants becoming consistent. Once the convergence conditions are met, the currently found optimal set of trajectory control node parameters will be output.

[0045] Through the above technical solution, this embodiment of the invention dynamically updates the pheromone matrix, enabling the algorithm to efficiently learn and reflect the quality of the search path, thus solving the challenge of how to efficiently update pheromones to dynamically reflect the quality of the search path. The pheromone matrix update mechanism, for example, combining pheromone volatilization and deposition, ensures that the search process can effectively utilize historical experience while avoiding getting trapped in local optima. Based on this, iterative search is performed using the updated pheromone matrix, and explicit convergence conditions are set, enabling the entire optimization process to intelligently converge to the target parameter set that meets the predetermined conditions. This not only improves search efficiency and avoids invalid iterations but also ensures the reliability and practicality of the output target trajectory control node parameter set. Combining the previous steps of constructing a continuous process parameter space based on the characteristic data of Chinese herbal raw materials, establishing a process trajectory function, and simulating extraction to generate multi-objective fitness values, this technical solution provides an adaptive and intelligent method for optimizing the extraction process of Chinese herbal components. It can dynamically adjust extraction parameters for different batches of Chinese herbal raw materials, thereby ensuring high yield while effectively controlling impurity dissolution and energy consumption, achieving multi-objective balance, and significantly improving the intelligence and refinement level of Chinese herbal extraction.

[0046] In the above embodiments of the present invention, the present invention proposes to update the pheromone matrix based on the multi-objective fitness values ​​of individual ant colonies to optimize the iterative search process. However, in this process, there may be a lack of effective mapping relationships and update mechanisms, resulting in inaccurate pheromone updates, affecting search efficiency and convergence, and making it difficult to achieve multi-objective balanced optimization.

[0047] To address this, this invention further proposes updating the pheromone matrix based on the multi-objective fitness values ​​of individual ant colonies to obtain an updated pheromone matrix. The specific steps include: establishing a mapping relationship between the pheromone matrix and control node parameters based on the multi-objective fitness values ​​of the individual ant colonies; determining the pheromone deposition amount for the corresponding control node parameters based on the established mapping relationship; performing pheromone volatilization processing on the pheromone matrix; and updating the pheromone matrix after pheromone volatilization processing based on the determined pheromone deposition amount to obtain the updated pheromone matrix.

[0048] Specifically, in establishing the mapping relationship between the pheromone matrix and the control node parameters, this mapping relationship aims to effectively connect the pheromone update mechanism in the ant colony algorithm with the actual optimization objective (multi-objective fitness value) and the parameters to be optimized (control node parameters). It transforms the abstract fitness value into guidance for specific control node parameters, ensuring that pheromone updates have practical meaning and directionality. One implementation method is to define a function or lookup table. For example, based on the quality of multi-objective fitness values ​​(such as Pareto front distance, objective function value, etc.), the fitness value range is divided into several levels, each level corresponding to a pheromone update weight or directly associated with a specific set of control node parameters. When an ant colony individual obtains a high multi-objective fitness value under a specific combination of control node parameters, this mapping relationship indicates that these control node parameters should obtain more pheromone. Another implementation method can be based on fuzzy logic or neural networks. For example, a neural network can be trained, with the multi-objective fitness value of an ant colony individual as input and the weights or influence factors related to each control node parameter in the pheromone matrix as output. In this way, nonlinear and more complex mapping relationships can be established to adapt to the complexity of multi-objective optimization problems in the extraction process of traditional Chinese medicine.

[0049] In determining the pheromone deposition amount for corresponding control node parameters, the pheromone deposition amount refers to the amount of pheromone added to the path (i.e., control node parameters) during the ant colony algorithm iteration process, based on the performance (multi-objective fitness value) of individual ants. Its role is to reinforce well-performing paths and guide subsequent ants to search towards better solutions. One implementation method is to calculate based on preset deposition rules. For example, for ants on the Pareto front, the control node parameters on their paths receive higher pheromone deposition amounts; for ants not on the Pareto front, the deposition amount is lower or zero. The deposition amount can be proportional to the fitness value, or a ranking mechanism can be used, with higher-ranked individuals receiving higher deposition amounts. Another implementation method is to use a method based on multi-objective decision analysis. For example, combining TOPSIS (Top-Solution Approximation) or DEA (Data Envelopment Analysis) methods, the multi-objective fitness values ​​of individual ants can be comprehensively evaluated to obtain a comprehensive score, which is then used as the basis for calculating the pheromone deposition amount. This allows for a more comprehensive consideration of the trade-offs between multiple objectives, determining a more reasonable pheromone deposition amount.

[0050] Furthermore, the pheromone evaporation process, a crucial mechanism in ant colony optimization (ACO), simulates the natural dissipation of pheromones over time. Its purpose is to prevent excessive pheromone accumulation on certain paths, thus avoiding premature convergence to local optima and enhancing the algorithm's exploration capabilities and adaptability to environmental changes. One implementation involves using a fixed evaporation rate. For example, after each iteration, all elements in the pheromone matrix are multiplied by a evaporation factor less than 1 (1-ρ), where ρ is the evaporation rate. This gradually reduces older pheromones, making room for new pheromone deposition. Another implementation uses a dynamic evaporation rate. For instance, the evaporation rate can be dynamically adjusted based on the algorithm's convergence or the number of iterations. In the early stages of the algorithm, a lower evaporation rate can be used to promote exploration; in the later stages, the evaporation rate can be appropriately increased to accelerate convergence, or adjusted based on the uniformity of pheromone distribution to maintain population diversity.

[0051] Furthermore, the final step in pheromone updating is updating the pheromone matrix after pheromone evaporation based on the determined pheromone deposition amount. This involves combining the evaporated pheromone with the newly deposited pheromone to form a new pheromone matrix. Its function is to dynamically adjust the attractiveness of each path in the search space, guiding subsequent ant colony individuals to search for optimal solutions more effectively. One implementation method is to use a simple additive update rule. That is, after pheromone evaporation, the pheromone deposition amount corresponding to each control node parameter is directly added to its corresponding position in the pheromone matrix. Another implementation method is to use a weighted average or more complex update rule. For example, while considering the deposition amount, a learning factor can be introduced so that the influence of newly deposited pheromones on the pheromone matrix can be adjusted. Alternatively, the deposition amount can be weighted based on the historical performance and current fitness value of the control node parameters to achieve a more refined update.

[0052] Through the above technical solution, this invention optimizes the pheromone update mechanism in the iterative search process of the ant colony algorithm. First, by establishing a mapping relationship between the pheromone matrix and control node parameters, the multi-objective fitness values ​​of individual ants are associated with specific control node parameters, making pheromone updates no longer blind but based on actual optimization results. This effectively solves the problem of the lack of an effective mapping relationship in traditional methods, ensuring the accuracy and directionality of pheromone updates. Second, based on this mapping relationship, the pheromone deposition amount of the corresponding control node parameters is accurately determined, allowing high-performing control node parameters to receive more reasonable and precise reinforcement, avoiding reliance on experience and subjective bias. Simultaneously, pheromone evaporation processing is applied to the pheromone matrix to effectively prevent excessive accumulation of pheromones on local paths, maintaining the algorithm's exploration ability and diversity, and avoiding premature entrapment in local optima. Finally, by combining evaporation processing and precise deposition amount updates to the pheromone matrix, dynamic and accurate iterative updates of pheromones are achieved. This enables the ant colony algorithm to perform iterative searches more effectively in the continuous process parameter space, significantly improving search efficiency and convergence stability. As a result, it can better achieve multi-objective balance optimization of the extraction process of traditional Chinese medicine components, and solve the problems of inaccurate updates, which affect search efficiency and convergence.

[0053] In the above embodiments of the present invention, the present invention also proposes to use the updated pheromone matrix for iterative search to optimize process parameters. However, in its implementation, how to establish effective selection rules to guide the search direction and determine the convergence condition in a timely manner during the search process in order to avoid invalid iterations and ensure search efficiency is a problem that needs to be solved.

[0054] To address this, this embodiment of the invention further proposes to utilize an updated pheromone matrix to iteratively search for trajectory control node parameters in a continuous process parameter space, outputting a target trajectory control node parameter set that satisfies convergence conditions. Specifically, this includes: establishing a transition selection rule based on the updated pheromone matrix; iteratively searching for the trajectory control node parameters in a continuous process parameter space according to the established transition selection rule; determining convergence conditions during the iterative search process; and outputting the target trajectory control node parameter set when the convergence conditions are met.

[0055] Establishing transition selection rules, based on the principles of the ant colony algorithm, uses an updated pheromone matrix to guide the movement of individual ants within a continuous process parameter space. This rule defines the probability of an ant choosing the next node from its current node, typically considering both pheromone concentration and heuristic information (e.g., objective function value or distance) along the path. For example, a roulette wheel selection (RWS) method can be used, where the probability of each optional path being selected is proportional to the product of its pheromone concentration and heuristic information; alternatively, a ranking-based selection method can be used, ranking all optional paths according to pheromone concentration and heuristic information, and assigning higher-ranked paths a higher probability of selection. By establishing transition selection rules, historical search experience can be effectively utilized, avoiding blind exploration and thus improving search efficiency and convergence speed.

[0056] Furthermore, based on the established transfer selection rules, the iterative search of the trajectory control node parameters in the continuous process parameter space refers to each individual in the ant colony gradually constructing its process trajectory in the continuous process parameter space according to the aforementioned transfer selection rules. Each ant starts from a starting point and selects the next trajectory control node parameter according to the transfer selection rules until a complete process trajectory is completed. This process is iterative; in each iteration, the ant colony individuals adjust their search path based on the current pheromone distribution and heuristic information. For example, multiple ant colony individuals can be simulated to explore the parameter space in parallel, with each individual independently selecting a path according to the rules and updating its pheromone after completing path construction. Alternatively, a phased search strategy can be adopted, first conducting a wide-area exploration, and then performing a local fine-grained search based on the preliminary results.

[0057] Here, determining convergence during the iterative search process refers to real-time monitoring of the ant colony algorithm's running status during its iterative optimization to determine whether the expected optimization objective or search stopping condition has been reached. Convergence conditions can be set based on various metrics. For example, convergence can be determined when the fitness value of the optimal solution no longer changes significantly over several consecutive iterations; or when the distribution of the pheromone matrix tends to stabilize, i.e., the rate of change of pheromone concentration on each path is below a certain threshold. Furthermore, a maximum number of iterations can be set as a convergence condition to prevent the algorithm from falling into an infinite loop. For example, convergence can be determined when the improvement in the global optimal fitness value is less than 0.01% over 10 consecutive iterations; or when the positions of all ant colonies in the parameter space are concentrated within a small range.

[0058] Specifically, when the convergence condition is met, the set of target trajectory control node parameters is output. This means that once the iterative search process meets the preset convergence condition, the algorithm will stop running and output the currently found optimal or a set of optimal trajectory control node parameters. These parameters represent the optimized trajectory of the extraction process of traditional Chinese medicine components and can be directly used to guide the actual extraction operation. For example, the set of trajectory control node parameters corresponding to the ant colony individual with the best multi-objective fitness value in all iterations can be output; or, if there are multiple non-dominated solutions (Pareto optimal solutions) that satisfy the optimization objective, the set of trajectory control node parameters composed of these non-dominated solutions can be output.

[0059] Through the above technical solutions, the embodiments of the present invention can effectively solve the problems of unclear selection rules and untimely convergence determination during iterative search. By establishing transition selection rules based on the updated pheromone matrix, the search behavior of individual ant colonies in the continuous process parameter space has clear guidance, enabling intelligent utilization of historical optimization information and avoiding blind searches caused by reliance on traditional experience, thereby significantly improving search efficiency. Simultaneously, the introduction of convergence condition determination during the iterative search process allows for real-time monitoring of the optimization status, timely identification of search stability points, and effective prevention of ineffective iterations and resource waste. When the convergence condition is met, the output target trajectory control node parameter set is fully optimized and reliable, and can be directly applied to the extraction process of traditional Chinese medicine components, ensuring the practicality and effectiveness of the optimization results. The synergistic effect of this series of technical features makes the optimization process of traditional Chinese medicine component extraction parameters more efficient, intelligent, and reliable, thus better addressing the challenges brought by batch differences in traditional Chinese medicine raw materials and achieving a stable improvement in extraction efficiency and product quality.

[0060] Secondly, traditional Chinese medicine extraction methods often employ fixed process parameters, ignoring the natural differences between batches of medicinal materials in terms of moisture content, cellulose content, tissue density, and initial content of the target active ingredient. This leads to unstable extraction efficiency and product quality, making it difficult to meet the requirements of modern pharmaceutical production standards. Existing extraction process optimization methods largely rely on empirical rules or offline experimental design, making it difficult to dynamically respond to changes in raw materials. Furthermore, traditional automated control systems lack the intelligent decision-making capabilities to handle the complex multi-objective and multi-constraint systems within the Chinese medicine framework. The pursuit of high yields often comes at the cost of high impurity dissolution and high energy consumption, making it difficult to achieve multi-objective balance in the process. Parameter adjustments are highly dependent on manual labor and experience, and the production process is slow to respond to sudden changes, hindering intelligent and refined production control.

[0061] Based on the above-mentioned deficiencies, this invention proposes a method for extracting components from traditional Chinese medicine, such as... Figure 2 As shown, the process includes the following steps: S201, converting the trajectory control node parameter set into a control instruction set for the extraction of traditional Chinese medicine components, the trajectory control node parameter set being output by the aforementioned process optimization method for realizing the extraction of traditional Chinese medicine components; S202, collecting the extraction conditions corresponding to the traditional Chinese medicine raw materials used for component extraction; S203, executing the control instruction set according to the extraction conditions to extract the target components.

[0062] In this embodiment of the invention, the trajectory control node parameter set is dynamically converted into a control instruction set, and the instructions are executed in conjunction with the real-time collected extraction conditions of the traditional Chinese medicine raw materials. This achieves intelligent response to batch-to-batch differences in raw materials and dynamic optimization of multi-objective process parameters. Specifically, in the implementation process, firstly, the trajectory control node parameter set output by the process optimization method is converted into an executable control instruction set. This conversion process ensures that the optimized parameters are accurately mapped to the control signals of the extraction equipment, such as the operating instructions of the temperature adjustment unit, solvent concentration control unit, and stirring speed control unit. Secondly, the extraction conditions of the traditional Chinese medicine raw materials are collected in real time through sensors or a preset database, including characteristic data such as the moisture content, cellulose content, tissue density, and initial content of the target effective components. These data directly reflect batch-to-batch differences. Finally, the control instruction set is dynamically adjusted and executed based on the collected extraction conditions, enabling the extraction process to adaptively respond to changes in raw materials and achieve precise control of parameters such as temperature, solvent concentration, and stirring speed.

[0063] Through the above technical solution, the embodiments of the present invention effectively solve the problems of low efficiency and unstable quality caused by batch variations of raw materials in traditional methods. Since the control instruction set is generated based on the trajectory control node parameter set output by the optimization method and executed in conjunction with real-time extraction conditions, the limitations of fixed parameter settings are avoided, significantly improving the response speed to changes in raw materials. Simultaneously, this method comprehensively considers multiple objective constraints such as yield, impurity content, and energy consumption during execution, achieving balanced optimization through dynamic adjustment of process parameters. This reduces reliance on manual experience, improves the intelligence level and refined control capability of the production process, and provides a reliable guarantee for the stability and efficiency of traditional Chinese medicine extraction processes.

[0064] Thirdly, traditional Chinese medicine extraction methods typically employ fixed process parameters (such as uniform temperature, solvent concentration, and stirring speed). This "one-size-fits-all" control approach ignores the natural differences between batches of medicinal materials in terms of moisture content, cellulose content, tissue density, and initial content of the target active ingredient, leading to unstable extraction efficiency and product quality, making it difficult to meet the requirements of modern pharmaceutical production standards. Existing extraction process optimization methods largely rely on empirical rules or offline experimental design, making it difficult to dynamically respond to changes in raw materials. Furthermore, traditional automated control systems lack the intelligent decision-making capabilities for the complex systems of Chinese medicine with multiple objectives and constraints. The pursuit of high yields often comes at the cost of high impurity dissolution and high energy consumption, making it difficult to achieve a balance between multiple objectives. Adjusting process parameters is highly dependent on manual labor and experience, resulting in a delayed response to sudden changes in the production process, hindering the achievement of intelligent and refined production control.

[0065] To address the aforementioned shortcomings, embodiments of the present invention provide a process optimization device for extracting components from traditional Chinese medicine, such as... Figure 3 As shown, the process optimization device 30 includes: a space construction module 301, used to construct a continuous process parameter space based on the characteristic data of Chinese herbal raw materials; a function establishment module 302, used to establish a process trajectory function based on the continuous process parameter space; a simulation extraction module 303, used to perform simulation extraction on ant colony individuals based on the trajectory control node parameters corresponding to the ant colony individuals, and generate the multi-objective fitness value of the corresponding ant colony individuals; and an iterative search module 304, used to perform iterative search on the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of the ant colony individuals, and output a set of target trajectory control node parameters.

[0066] In this embodiment of the invention, by combining the space construction module 301 and the function establishment module 302 in a dynamic modeling manner, and introducing the collaborative mechanism of the simulated extraction module 303 and the iterative search module 304, a personalized process parameter space is constructed for batch differences of Chinese medicinal materials and multi-objective intelligent optimization is achieved, so as to achieve a balance between ensuring high yield, suppressing impurity dissolution and reducing energy consumption.

[0067] In one embodiment, the space construction module 301 is specifically used to collect characteristic data of Chinese herbal raw materials; determine the solid-liquid ratio of the Chinese herbal raw materials based on the collected characteristic data; read the capacity boundary parameters of the extraction equipment to form a set of equipment capacity boundaries; and construct a continuous process parameter space by combining the feature vector of the Chinese herbal raw materials and the obtained solid-liquid ratio under the constraints of the set of equipment capacity boundaries.

[0068] Specifically, the space construction module 301 directly constructs a continuous process parameter space based on the characteristic data of the Chinese herbal medicine raw materials. This characteristic data includes quantitative indicators such as the moisture content, cellulose content, tissue density, and initial content of the target active ingredient, used to characterize the inherent differences between different batches of raw materials. Since the raw materials of Chinese herbal medicine exhibit natural fluctuations, the continuous process parameter space generated by this module can cover the feasible domain of process parameters such as temperature, solvent concentration, stirring speed, and extraction time, thereby avoiding the limitations of fixed parameter control and providing a personalized basis for subsequent optimization.

[0069] Here, the function establishment module 302 establishes a process trajectory function based on the aforementioned continuous process parameter space. This function defines the dynamic path of process parameters changing over time during the extraction process. Specifically, the process trajectory function describes the continuous evolution of parameters on the time axis through a mathematical model, such as the smooth transition of temperature from an initial value to a target value, rather than using a stepped fixed value. Given the nonlinear dynamic characteristics of the traditional Chinese medicine extraction process, this module can generate a dynamic trajectory that adapts to the characteristics of the raw materials, replacing the static settings of traditional empirical rules, thereby achieving real-time response capability to changes in raw materials.

[0070] In one embodiment, the simulation extraction module 303 is specifically used to read the trajectory control node parameters corresponding to individual ant colonies to form a set of trajectory control node parameters; generate the process trajectory of individual ant colonies within the extraction time range based on the formed set of trajectory control node parameters; use the generated process trajectory of individual ant colonies as input boundary conditions, and perform simulation extraction on the individual ant colonies in a model simulation environment to generate the multi-objective fitness value corresponding to the individual ant colonies.

[0071] Specifically, the simulated extraction module 303 performs simulated extraction on individual ants based on the trajectory control node parameters corresponding to each individual ant, generating a multi-objective fitness value for each individual ant. Each individual ant represents a potential process scheme, and its trajectory control node parameters define the key control points of the process trajectory function. During the simulation, the module calls a preset extraction kinetic model, inputs the characteristic data of the Chinese herbal raw materials and the trajectory control node parameters, predicts indicators such as the yield of target components, impurity content, and energy consumption, and generates a multi-objective fitness value through weighted aggregation. Considering the mutual constraints in multi-objective optimization, this module can quantitatively evaluate the comprehensive performance of the scheme in terms of yield, purity, and energy efficiency, effectively resolving the contradiction between high yield and high impurity dissolution in traditional methods.

[0072] In one embodiment, the iterative search module 304 is specifically used to update the pheromone matrix according to the multi-objective fitness values ​​of the ant colony individuals to obtain the updated pheromone matrix; and to use the updated pheromone matrix to iteratively search the trajectory control node parameters in the continuous process parameter space to output a set of target trajectory control node parameters that meet the convergence conditions.

[0073] Specifically, the iterative search module 304 iteratively searches for trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of individual ant colonies, outputting a set of target trajectory control node parameters. This module employs an improved ant colony algorithm, guiding the search direction based on multi-objective fitness values: ant colonies with higher fitness values ​​release higher pheromone concentrations, thus attracting subsequent searches to converge towards high-quality areas. Simultaneously, to avoid local optima, the algorithm introduces a random perturbation mechanism, dynamically adjusting the search step size in each iteration. Because the iterative process continuously optimizes the trajectory control node parameters, it can efficiently converge to the globally optimal solution that satisfies multi-objective constraints. The final output set of target trajectory control node parameters can directly drive the operation of the actual extraction equipment.

[0074] Through the above technical solutions, the embodiments of the present invention effectively realize intelligent and refined control of the extraction process of traditional Chinese medicine. The spatial construction module constructs a personalized parameter space based on raw material differences, the function establishment module generates a dynamic process trajectory, the simulated extraction module quantitatively evaluates multi-objective performance, and the iterative search module intelligently optimizes parameter combinations. Overall, this device effectively overcomes the reliance on manual experience in traditional methods, significantly improves the dynamic response capability to raw material fluctuations, and while ensuring a high yield of target components, suppresses impurity dissolution and reduces energy consumption, meeting the multi-objective balance requirements of modern pharmaceutical production.

[0075] Fourthly, traditional Chinese medicine extraction methods typically employ fixed process parameters, ignoring the natural differences between batches of medicinal materials in terms of moisture content, cellulose content, tissue density, and initial content of the target active ingredient. This leads to unstable extraction efficiency and product quality, making it difficult to meet the requirements of modern pharmaceutical production standards. Existing extraction process optimization methods largely rely on empirical rules or offline experimental design, making it difficult to dynamically respond to changes in raw materials. Furthermore, traditional automated control systems lack the intelligent decision-making capabilities to handle multiple objectives and constraints within the complex systems of Chinese medicine. The pursuit of high yields often comes at the cost of high impurity dissolution and high energy consumption, making it difficult to achieve a balance between multiple objectives. Adjusting process parameters is highly dependent on manual labor and experience, resulting in a delayed response to sudden changes and hindering intelligent and refined production control.

[0076] To address the aforementioned shortcomings, embodiments of the present invention provide a device for extracting components from traditional Chinese medicine, such as... Figure 4 As shown, the device 40 includes: a conversion module 401, used to convert the trajectory control node parameter set into a control instruction set for the extraction of traditional Chinese medicine components, the trajectory control node parameter set being output by the aforementioned process optimization device for realizing the extraction of traditional Chinese medicine components; a data acquisition module 402, used to acquire the extraction conditions corresponding to the traditional Chinese medicine raw materials used for component extraction; and a component extraction module 403, used to execute the control instruction set according to the extraction conditions to extract the target components.

[0077] In this embodiment of the invention, by combining the conversion module 401, the acquisition module 402 and the component extraction module 403 in an orderly manner, and by combining the intelligent conversion of the trajectory control node parameter set with the real-time extraction condition acquisition in a dynamic and collaborative manner, the adaptive response to batch differences of Chinese herbal raw materials and the real-time optimization of multi-objective process parameters can be effectively achieved.

[0078] Specifically, the conversion module 401 converts the set of trajectory control node parameters output by the aforementioned process optimization device into an operable set of control instructions, avoiding the limitations of rules of thumb and ensuring the adaptability of the instructions; the acquisition module 402 acquires characteristic data of the Chinese herbal raw materials in real time, such as moisture content and cellulose content, directly responding to the natural changes of the raw materials and solving the problem of unstable efficiency caused by ignoring batch differences; the component extraction module 403 executes the control instruction set according to the acquired extraction conditions, combining real-time conditions and optimized parameters to effectively balance multiple objectives such as yield, impurity dissolution, and energy consumption. Since the set of trajectory control node parameters processed by the conversion module 401 is dynamically generated based on an intelligent optimization algorithm, it can adapt to different raw material characteristics, thus significantly reducing impurity dissolution and energy consumption while achieving high yield; at the same time, the real-time data input of the acquisition module 402 shortens the response time of the device to changes in raw materials, overcoming the lag of manual adjustment.

[0079] Through the above technical solutions, the embodiments of the present invention effectively solve the problems of fixed parameters, reliance on experience, inability to dynamically respond to changes in raw materials, response lag, and difficulty in achieving multi-objective balance in traditional Chinese medicine extraction methods. Overall, the device improves the intelligence level and refined control capability of the extraction process through modular collaborative operation, ensures the stability of product quality, and reduces resource consumption in the production process.

[0080] Fifthly, embodiments of the present invention provide a computer-readable storage medium. This computer-readable storage medium stores computer-executable instructions, which, when invoked or executed by a processor, cause the processor to implement a process optimization method for the extraction of traditional Chinese medicine components. Specifically, the computer-executable instructions encode the core logic of the optimization algorithm, including constructing a continuous process parameter space based on the characteristic data of the traditional Chinese medicine raw materials, establishing a process trajectory function based on the continuous process parameter space, simulating extraction of individual ants based on the trajectory control node parameters corresponding to individual ants to generate multi-objective fitness values, and iteratively searching the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of individual ants and outputting a set of target trajectory control node parameters.

[0081] Similarly, embodiments of the present invention also provide a computer-readable storage medium. This computer-readable storage medium stores computer-executable instructions, which, when invoked or executed by a processor, cause the processor to implement a method for extracting traditional Chinese medicine components. Specifically, the computer-executable instructions encode the core logic of an optimization algorithm, including converting a set of trajectory control node parameters into a set of control instructions for extracting traditional Chinese medicine components. The set of trajectory control node parameters is obtained from the output of the process optimization method for extracting traditional Chinese medicine components as described in any of the above embodiments; collecting the extraction conditions corresponding to the traditional Chinese medicine raw materials used for component extraction; and executing the set of control instructions according to the extraction conditions to extract the target component.

[0082] Through the above technical solution, the computer storage medium serves as the physical carrier, enabling the optimization method to be persistently stored and deployed on different devices, thus solving the problem of reliance on human experience. The computer-executable instructions cause the processor to execute the optimization process in real time, dynamically adjusting the trajectory control node parameters based on the input herbal raw material characteristic data and equipment boundary conditions, and performing multi-objective fitness evaluation and iterative search. Because the optimization algorithm can comprehensively consider multiple objective constraints such as maximizing yield, minimizing impurities, and minimizing energy consumption, it generates a dynamic process trajectory adapted to the characteristics of the raw materials in a continuous process parameter space, thereby achieving intelligent decision-making for complex herbal systems. Compared to traditional fixed-parameter control methods, this application effectively overcomes the process instability problem caused by batch-to-batch differences in raw materials through the deployment and execution of digital instructions, while avoiding the contradiction between high impurity dissolution and high energy consumption. It improves extraction efficiency while ensuring product quality stability, ultimately achieving the goal of intelligent and refined production control.

[0083] The following will provide a more detailed explanation of the above-mentioned technical solution of the present invention through a more specific example: A pharmaceutical company plans to extract active ingredients from a specific batch of traditional Chinese medicine raw materials. The moisture content, cellulose content, and initial concentration of the target active ingredient in this batch of raw materials differ from historical batches. Traditional extraction methods with fixed process parameters are ill-suited to these batch-to-batch variations, leading to fluctuations in extraction efficiency and product quality. To address this issue, the company has adopted a process optimization method.

[0084] First, the system collects characteristic data for this batch of Chinese herbal medicine raw materials, including their moisture content, cellulose content, tissue density, and the initial content of the target active ingredient. Based on this characteristic data, the system determines the solid-liquid ratio range for this batch of raw materials, for example, setting it between 1:8 and 1:12. Simultaneously, the system reads the capability boundary parameters of existing extraction equipment, such as the maximum extraction temperature of 95℃, the minimum extraction temperature of 60℃, the maximum stirring speed of 200 rpm, the minimum stirring speed of 50 rpm, the solvent concentration range of 50% to 80% ethanol, and the maximum extraction time of 3 hours. Under the constraints of these equipment capability boundaries, combined with the characteristic vector of the Chinese herbal medicine raw materials and the determined solid-liquid ratio, the system constructs a continuous process parameter space. This space defines the variable ranges of key parameters such as extraction temperature, solvent concentration, stirring speed, and extraction time; for example, the extraction temperature range is refined to 70℃ to 85℃, and the solvent concentration range is refined to 60% to 75% ethanol, to adapt to the characteristics of this batch of raw materials.

[0085] Next, based on the constructed continuous process parameter space, the system establishes a process trajectory function. This function can describe how key process parameters (such as temperature and solvent concentration) dynamically change over time throughout the extraction process, thereby achieving fine-grained control of the extraction process rather than using a single fixed value. For example, the function can set the temperature to gradually increase from 70°C to 80°C in the initial stage of extraction and then remain stable in the subsequent stages.

[0086] Subsequently, the system initiates the ant colony optimization algorithm. Each "ant colony individual" represents a set of potential trajectory control node parameters, which define the key points of a specific process trajectory. The system reads the trajectory control node parameters corresponding to each ant colony individual, forming a set of trajectory control node parameters. For example, an ant colony individual might correspond to control nodes with a temperature of 70°C and a solvent concentration of 65% at extraction time 0 minutes, 80°C and 65% at extraction time 60 minutes, and 80°C and 70% at extraction time 120 minutes. Based on these node parameters, the system generates a continuous process trajectory for the ant colony individual over the entire extraction time range (e.g., 180 minutes). Then, this generated process trajectory is used as input boundary conditions to simulate extraction of the ant colony individual in a validated model simulation environment. The simulation results generate multi-objective fitness values ​​corresponding to this ant colony individual, such as a target component yield of 92%, impurity content of 5%, and energy consumption of 150 kWh. These values ​​comprehensively reflect the performance of the process trajectory.

[0087] After obtaining the multi-objective fitness values ​​of individual ant colonies, the system iteratively searches for trajectory control node parameters in a continuous process parameter space based on these values. Specifically, the system updates the pheromone matrix based on the multi-objective fitness values ​​of individual ant colonies. This update process includes: first, establishing a mapping relationship between the pheromone matrix and control node parameters; second, based on this mapping relationship, determining the pheromone deposition amount corresponding to the control node parameters traversed by high-performing ant colonies; third, performing pheromone evaporation processing on the pheromone matrix to avoid local optima; and finally, updating the pheromone matrix after pheromone evaporation processing based on the determined pheromone deposition amount. Using the updated pheromone matrix, the system establishes a transition selection rule to guide new ant colony individuals in selecting trajectory control node parameters within a continuous process parameter space. New ant colony individuals tend to choose paths with higher pheromone concentrations, i.e., process parameter combinations that have shown good fitness values ​​in previous iterations. Based on this transition selection rule, the system iteratively searches for trajectory control node parameters within the continuous process parameter space. During the iterative search, the system continuously determines convergence conditions. For example, convergence is considered satisfied when the optimal fitness value no longer significantly improves over multiple generations, or when the preset maximum number of iterations is reached. Once convergence is satisfied, the system outputs the final set of target trajectory control node parameters. This set represents the optimal dynamic extraction process for this batch of Chinese herbal medicine raw materials, considering multiple objectives such as yield, purity, and energy consumption.

[0088] Compared to traditional methods, this invention no longer relies on fixed empirical parameters, but dynamically constructs and optimizes a parameter space based on the unique characteristics of each batch of Chinese herbal raw materials. Through ant colony optimization and model simulation, it achieves balanced optimization of multiple objectives (such as high yield, low impurities, and low energy consumption), overcoming the lack of intelligent decision-making capabilities in traditional methods when dealing with complex systems of Chinese herbal medicine. This dynamic and intelligent optimization process significantly improves the adaptability and robustness of the extraction process, ensuring product quality stability and production efficiency, and effectively solving the problems of traditional process parameter adjustments being highly dependent on manual labor and experience, and lagging response to sudden changes.

[0089] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, modifications can still be made to the embodiments. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A process optimization method for extracting components from traditional Chinese medicine, characterized in that, The method includes: A continuous process parameter space is constructed based on the characteristic data of Chinese medicinal raw materials; A process trajectory function is established based on the continuous process parameter space; Based on the trajectory control node parameters corresponding to each ant colony individual, the individual ant colony is simulated and extracted to generate a multi-target fitness value corresponding to each ant colony individual; Based on the multi-objective fitness values ​​of individual ant colonies, the trajectory control node parameters are iteratively searched in the continuous process parameter space to output a set of target trajectory control node parameters.

2. The process optimization method according to claim 1, characterized in that, The construction of a continuous process parameter space based on the characteristic data of traditional Chinese medicine raw materials includes: Collect characteristic data of Chinese medicinal materials; Based on the collected characteristic data of the Chinese herbal raw materials, the solid-liquid ratio of the Chinese herbal raw materials is determined; Read the capacity boundary parameters of the extraction equipment to form a set of equipment capacity boundaries; Under the constraints of the set of equipment capability boundaries, a continuous process parameter space is constructed by combining the feature vectors of the Chinese herbal raw materials and the obtained solid-liquid ratio.

3. The process optimization method according to claim 1, characterized in that, The step of simulating and extracting the ant colony individuals based on the trajectory control node parameters corresponding to the individual ant colony individuals, and generating multi-target fitness values ​​corresponding to the individual ant colony individuals, includes: Read the trajectory control node parameters corresponding to each individual ant in the colony and form a set of trajectory control node parameters; Based on the set of trajectory control node parameters formed, the process trajectory of individual ant colonies is generated within the extraction time range. The process trajectory of the generated ant colony individuals is used as the input boundary condition. The ant colony individuals are simulated and extracted in the model simulation environment to generate the multi-objective fitness value corresponding to the ant colony individuals.

4. The process optimization method according to claim 1, characterized in that, Based on the multi-objective fitness values ​​of individual ant colonies, the trajectory control node parameters are iteratively searched in the continuous process parameter space to output a set of target trajectory control node parameters, including: The pheromone matrix is ​​updated based on the multi-objective fitness values ​​of the individual ants in the ant colony, resulting in the updated pheromone matrix. Using the updated pheromone matrix, the trajectory control node parameters are iteratively searched in the continuous process parameter space to output a set of target trajectory control node parameters that meet the convergence condition.

5. The process optimization method according to claim 4, characterized in that, The pheromone matrix is ​​updated based on the multi-objective fitness values ​​of the individual ant colonies, resulting in the updated pheromone matrix, which includes: Based on the multi-objective fitness values ​​of individual ant colonies, a mapping relationship is established between the pheromone matrix and the control node parameters; Based on the established mapping relationship between the pheromone matrix and the control node parameters, the pheromone deposition amount of the corresponding control node parameters is determined. The pheromone matrix is ​​subjected to pheromone evaporation treatment; Based on the determined pheromone deposition amount, the pheromone matrix after pheromone volatilization treatment is updated to obtain the updated pheromone matrix.

6. The process optimization method according to claim 4, characterized in that, The updated pheromone matrix is ​​used to iteratively search the trajectory control node parameters in the continuous process parameter space, outputting a set of target trajectory control node parameters that satisfy the convergence condition, including: Based on the updated pheromone matrix, a transition selection rule is established; Based on the established transfer selection rules, the trajectory control node parameters are iteratively searched in the continuous process parameter space; During the iterative search process, convergence conditions are determined. When the convergence condition is met, the set of parameters for the target trajectory control nodes is output.

7. A method for extracting components from traditional Chinese medicine, characterized in that, The method includes: The trajectory control node parameter set is converted into a control instruction set for the extraction of traditional Chinese medicine components, wherein the trajectory control node parameter set is output by the process optimization method for realizing the extraction of traditional Chinese medicine components as described in any one of claims 1 to 6; Extraction conditions corresponding to the Chinese herbal raw materials used for component extraction were collected; Based on the extraction conditions, the control instruction set is executed to extract the target component.

8. A process optimization device for extracting components from traditional Chinese medicine, characterized in that, The device includes: The spatial construction module is used to construct a continuous process parameter space based on the characteristic data of Chinese herbal raw materials; The function creation module is used to create a process trajectory function based on the continuous process parameter space; The simulation extraction module is used to simulate and extract the individual ant colony based on the trajectory control node parameters corresponding to the individual ant colony, and generate a multi-target fitness value corresponding to the individual ant colony. The iterative search module is used to iteratively search the trajectory control node parameters in the continuous process parameter space based on the multi-objective fitness values ​​of the individual ant colony members, and output the target trajectory control node parameter set.

9. A device for extracting components from traditional Chinese medicine, characterized in that, The device includes: A conversion module is used to convert the trajectory control node parameter set into a control instruction set for the extraction of traditional Chinese medicine components, wherein the trajectory control node parameter set is output by the process optimization device for realizing the extraction of traditional Chinese medicine components as described in claim 8; The data acquisition module is used to acquire the extraction conditions corresponding to the Chinese herbal raw materials used for component extraction. The component extraction module is used to execute the control instruction set according to the extraction conditions to extract the target component.

10. A computer storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked or executed by a processor, cause the processor to implement the method as described in any one of claims 1 to 7.

Images