Plant composite composition formula optimization method and plant composite composition

Abstract

The invention discloses a plant composite composition formula optimization method and a plant composite composition, and relates to the field of plant formulae, the method comprises the following steps: obtaining multi-source standardized plant characteristic data; modeling each plant raw material component in the multi-source standardized plant feature data to form a plurality of interacting subjects; an interaction rule and a global target are set, and a plant composite composition candidate formula is formed through a dynamic simulation cooperation process; constructing a multi-objective optimization framework including an effectiveness objective, a stability objective, a safety objective, a cost objective and a sustainability objective, and configuring constraint conditions; and based on a multi-objective optimization algorithm and a random forest algorithm, in combination with a multi-objective optimization framework and constraint conditions, sorting and screening the plant composite composition candidate formulas to obtain a plant composite composition formula set. According to the invention, the effectiveness, the stability, the safety, the cost and the sustainability can be comprehensively balanced, and a more scientific and easier high-quality formula is obtained.

Description

Technical Field

[0001] The present invention relates to the field of plant formulations, and more particularly, to a method for optimizing a plant composite composition formulation and a plant composite composition. Background Art

[0002] Plant composite compositions are widely used in multiple industries such as medicine, food, health products, and agriculture due to their natural activity, low toxicity, and multifunctionality. With the deepening of scientific research, more and more plant-derived active ingredients have been discovered and proven to have synergistic effects, and the overall effect can be significantly improved through reasonable proportioning. However, there are often non-linear interactions between different plant components, and combined with various factors such as raw material costs, processing technologies, safety, and regulatory compliance, the optimization and development of plant composite composition formulations have become a highly complex and systematic task.

[0003] The following problems also exist in the process of optimizing the plant composite composition formulation: (1) Traditional single-index or empirical methods are prone to missing component cooperation effects and various rigid constraints. How to accurately model the complex interactions, synergistic, and antagonistic relationships between various plant raw materials and components, and avoid the distortion of formulation performance caused by simple superposition or neglect of cooperation effects.

[0004] (2) How to automatically and efficiently search for and screen out high-quality combined formulations, improve development efficiency, and accelerate the transformation of achievements.

[0005] In response to the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0006] In response to the problems in the related art, the present invention provides a method for optimizing a plant composite composition formulation and a plant composite composition to overcome the above technical problems existing in the existing related art. [[ID=!25]]

[0007] To this end, the specific technical solutions adopted by the present invention are as follows: According to one aspect of the present invention, there is provided a method for optimizing a plant composite composition formulation, including: S1. Preprocessing multi-domain data of the plant composite composition to obtain multi-source standardized plant feature data; S2. Modeling each plant raw material component in the multi-source standardized plant feature data as interacting multi-agents; setting interaction rules and a global goal, and forming a candidate formulation of the plant composite composition through a dynamic simulation collaboration process; S3. Constructing a multi-objective optimization framework including effectiveness objectives, stability objectives, safety objectives, cost objectives, and sustainability objectives, and configuring constraint conditions; S4. Based on the multi-objective optimization algorithm and the random forest algorithm, combined with the multi-objective optimization framework and constraint conditions, sort and screen the candidate formulas of the plant composite composition to obtain the formula set of the plant composite composition.

[0008] Further, by preprocessing the multi-domain data of the plant composite composition, multi-source normalized plant feature data is obtained, including: Collect multi-domain data of the plant composite composition, including chemical composition data, biological activity data, toxicological data, processing data, and storage data of plant raw materials; Unify the formats of the multi-domain data of the plant composite composition, remove noise and outliers, fill in the missing values, and standardize the dimensions and units to form multi-source normalized plant feature data.

[0009] Further, in the multi-source normalized plant feature data, each plant raw material component is modeled as interacting multi-agents; set interaction rules and global goals, and through the dynamic simulation cooperation process, form the candidate formulas of the plant composite composition, including: In the multi-source normalized plant feature data, each plant raw material component is modeled as an independent agent and given the attributes of an independent agent; determine the preliminary behavior rules of the independent agent and initialize the ratios of each plant raw material component; By mining experimental data and literature, adjust the preliminary behavior rules of the independent agent and determine the interaction rules between the independent agents; introduce external environmental impacts and regulatory constraints, and set global goals; Establish a kinetic mapping relationship between the interaction rules of the independent agent and the global goal; set the number of iterations and simulation parameters, and in each iteration, each independent agent autonomously adjusts the ratio and action according to the interaction rules and feedback, simulating the real dynamic cooperation process; Record the formulas of the plant composite composition in each iteration in real time, and comprehensively form the candidate formulas of the plant composite composition.

[0010] Further, determining the preliminary behavior rules of the independent agent and initializing the ratios of each plant raw material component includes: For each set of independent agent attribute parameters, combined with literature reports and industry common sense, set the preliminary behavior rules of the independent agent; According to the basic goals of the plant composite composition design, use the normalized allocation method to determine the ratios of each plant raw material component.

[0011] Further, by mining experimental data and literature, adjust the preliminary behavior rules of the independent agent and determine the interaction rules between the independent agents; introduce external environmental impacts and regulatory constraints, and set global goals, including: Obtain a structured set of biological interaction information related to independent entities and a normalized literature evidence entry library, compare them with the preliminary behavior rules of the independent entities, dynamically adjust the boundary of the behavior rules of the independent entities, and obtain the set of behavior rules of the independent entities; Construct interaction rules between independent entities according to the set of behavior rules of the independent entities; Obtain and quantify the external environmental impact and regulatory constraint parameters, and use them as formulation optimization constraints throughout the process. On the basis of forming a coherent mapping with all the behavior rules of the independent entities, construct a global objective.

[0012] Furthermore, establish a dynamic mapping relationship between the interaction rules of independent entities and the global objective; set the number of iterations and simulation parameters, and in each iteration, each independent entity autonomously adjusts the ratio and actions according to the interaction rules and feedback, and simulate the real - world dynamic collaboration process, including: Analyze the interaction rules of independent entities and the global objective, and model the impact of the behavior changes and ratio adjustments of each independent entity on the change of the objective function as a causal dynamic mapping; Use differential equations to clarify the response relationships among behavior, state, and objective, and form a set of dynamic rules; Configure the simulation parameters and initialize the simulation environment. In each round of simulation iteration, all independent entities adjust their own ratios, attributes, and behavior states according to the dynamic mapping model, and calculate in real - time the impact of the adjustment results of each independent entity on the global objective; Each independent entity iterates according to its own state and external feedback information, and continuously evolves the dynamic coupling process between the behavior of the independent entity and the global objective; Save the historical state, formulation combination, and objective results of each iteration for the screening and analysis of the final candidate formulations.

[0013] Furthermore, construct a multi - objective optimization framework including effectiveness objectives, stability objectives, safety objectives, cost objectives, and sustainability objectives, and configure the constraint conditions including: According to the determined multi - objective system, configure mathematical indicators for the effectiveness objective, stability objective, safety objective, cost objective, and sustainability objective; Perform unit normalization on each objective indicator, and assign weights to different objectives to form a weighted objective function; Construct all constraint conditions, and solidify the constraint conditions into the input of the optimization algorithm in the form of operators.

[0014] Furthermore, based on the multi - objective optimization algorithm and the random forest algorithm, and combined with the multi - objective optimization framework and constraint conditions, rank and screen the candidate formulations of the plant composite composition to obtain the plant composite composition formulation set, including: Convert each candidate formula of the plant composite composition into a characteristic parameter vector corresponding to the multi-objective index system, and standardize and unify each numerical value; at the same time, filter out all candidate formulas of the plant composite composition that violate the constraint conditions; Based on the multi-objective optimization algorithm, comprehensively consider each objective of the multi-objective optimization framework, perform global optimization and sorting on the candidate formulas of the plant composite composition, and extract a subset of formulas in the optimal trade-off state on the multi-objective solution space; Use the random forest model, with the multi-dimensional characteristic parameters of the candidate formulas of the plant composite composition as the input and the composite performance indicators measured by experiments as the output, to perform supervised learning training; use the trained random forest model to predict and evaluate the performance of each formula in the formula subset, and obtain the comprehensive performance prediction scores corresponding to each formula subset; Arrange in descending order according to the comprehensive performance prediction scores, and output the plant composite composition formula set with the top comprehensive performance ranking and meeting the multi-objective constraint conditions.

[0015] Furthermore, based on the multi-objective optimization algorithm, comprehensively consider each objective of the multi-objective optimization framework, perform global optimization and sorting on the candidate formulas of the plant composite composition, and extracting a subset of formulas in the optimal trade-off state on the multi-objective solution space includes: Select a multi-objective optimization algorithm and initialize the population; run the main loop of the multi-objective optimization algorithm, and use population evolution to continuously evolve the candidate formulas in the multi-objective solution space; After several iterations, obtain the multi-objective solution set of the formulas optimized by multi-objective; According to the Pareto optimality principle, identify and extract the candidate formulas of the plant composite composition in the multi-objective solution set space of the formulas where the trade-off state of each objective is optimal and cannot be comprehensively dominated by other solutions, and form a subset of formulas in the optimal trade-off state.

[0016] According to another aspect of the present invention, there is also provided a plant composite composition, including: 1 - 1.8 parts of licorice; 0.4 - 1.2 parts of platycodon root; 0.4 - 0.6 parts of polygonatum odoratum; 0.2 - 1.0 parts of phyllanthus emblica; 0.8 - 1.0 parts of citri reticulatae pericarpium.

[0017] The beneficial effects of the present invention are: (1) The present invention can comprehensively balance effectiveness, stability, safety, cost, and sustainability to obtain a more scientific and more implementable high-quality formulation. Through kinetic multi-agent modeling and dynamic simulation, it can systematically explore the complex interaction laws between components, greatly increase the space for combinatorial innovation, and can automatically adapt and iterate according to environmental and ratio changes, with the simulation results having higher real-world reliability. Introducing a multi-objective global optimization algorithm can accelerate the search for the optimal formulation, and superimposing machine learning models can accurately predict performance, with the overall efficiency and accuracy being significantly superior to traditional human-machine iterative screening methods.

[0018] (2) Strengthen the constraint-driven optimization process to eliminate actual implementation obstacles such as ratio violations and risk overrun. The optimized formulation is highly controllable in dimensions such as regulations and safety. It can continuously incorporate new active ingredients, processes, and regulatory data to continuously support in-depth development and industrial application in different demand fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 is a flowchart of a method for optimizing the formulation of a plant composite composition according to an embodiment of the present invention; Figure 2 is a line graph showing the influence of the addition amount of licorice on the sensory score according to an embodiment of the present invention; Figure 3 is a line graph showing the influence of the addition amount of platycodon root on the sensory score according to an embodiment of the present invention; Figure 4 is a line graph showing the influence of the addition amount of phyllanthus emblica on the sensory score according to an embodiment of the present invention; Figure 5 is a line graph showing the influence of the addition amount of polygonatum odoratum on the sensory score according to an embodiment of the present invention; Figure 6 is a line graph showing the influence of the addition amount of citrus grandis 'Tomentosa' on the sensory score according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To further illustrate the embodiments, the present invention provides drawings. These drawings are a part of the disclosure of the present invention, mainly used to illustrate the embodiments, and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, those of ordinary skill in the art should be able to understand other possible implementation manners and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are usually used to represent similar components.

[0022] According to an embodiment of the present invention, a method for optimizing the formula of a plant composite composition and a plant composite composition are provided.

[0023] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. As Figure 1 shown, according to an embodiment of the present invention, a method for optimizing the formula of a plant composite composition is provided, including: S1. By preprocessing the multi-domain data of the plant composite composition, multi-source standardized plant characteristic data is obtained.

[0024] S2. In the multi-source standardized plant characteristic data, each plant raw material component is modeled as interacting multi-agents; interaction rules and global goals are set, and through the dynamic simulation cooperation process, a candidate formula for the plant composite composition is formed.

[0025] S3. A multi-objective optimization framework including effectiveness objectives, stability objectives, safety objectives, cost objectives, and sustainability objectives is constructed, and constraint conditions are configured.

[0026] S4. Based on the multi-objective optimization algorithm and the random forest algorithm, and in combination with the multi-objective optimization framework and constraint conditions, the candidate formulas for the plant composite composition are sorted and screened to obtain a set of formulas for the plant composite composition.

[0027] In one embodiment, by preprocessing the multi-domain data of the plant composite composition, the multi-source standardized plant characteristic data obtained includes: Collect the multi-domain data of the plant composite composition, including the chemical composition data, biological activity data, toxicological data, processing data, and storage data of the plant raw materials; unify the format of the multi-domain data of the plant composite composition, remove noise and outliers, fill in the missing values, and perform dimension and unit standardization to form multi-source standardized plant characteristic data.

[0028] The chemical composition data includes the specific composition, relative content, molecular structure information, etc. of various active ingredients in plant raw materials; the biological activity data includes experimental measurements, databases, literature reports, etc. For example: the effectiveness, functionality, and activity parameters obtained from experiments (such as free radical scavenging rate, changes in inflammatory factor levels, etc.), the information records of the active ingredients and their functional properties of plant raw materials in the database, and the data information on the physiological activities and efficacy characteristics of plants or ingredients clearly recorded in authoritative literature. The biological activity data is used to comprehensively reflect the activity performance and functional characteristics of plants and their combinations; the functional characteristics of various plant raw materials; the toxicological data includes acute and chronic toxicity, sensitization, teratogenicity, safety dose limits, etc. The processing data includes the parameters and characteristics of plant raw materials in the processes of extraction, concentration, drying, grading, mixing, packaging, etc. The storage data includes the data related to the shelf life, storage temperature and humidity, variability, and degradation rate of raw materials and finished products.

[0029] In one embodiment, in the multi-source standardized plant characteristic data, each plant raw material component is modeled as interacting multi-agents; interaction rules and a global goal are set, and through a dynamic simulation collaboration process, a candidate formula for the plant composite composition is formed, including: In the multi-source standardized plant characteristic data, each plant raw material component is modeled as an independent agent and given independent agent attributes; the preliminary behavior rules of the independent agents are determined, and the ratios of each plant raw material component are initialized; by mining experimental data and literature, the preliminary behavior rules of the independent agents are adjusted, and the interaction rules between the independent agents are determined; external environmental impacts and regulatory constraints are introduced, and a global goal is set; a kinetic mapping relationship between the interaction rules of the independent agents and the global goal is established; the number of iterations and simulation parameters are set, and in each iteration, each independent agent autonomously adjusts the ratio and actions according to the interaction rules and feedback, simulating the real dynamic collaboration process; the formula of the plant composite composition in each iteration is recorded in real time, and a candidate formula for the plant composite composition is comprehensively formed.

[0030] In one embodiment, determining the preliminary behavior rules of the independent agents and initializing the ratios of each plant raw material component includes: For each set of independent agent attribute parameters, and combined with literature reports and industry common sense, the preliminary behavior rules of the independent agents are set; according to the basic goal of the plant composite composition design (such as safety first or cost first), the normalized allocation method is used to determine the ratios of each plant raw material component.

[0031] In one embodiment, by mining experimental data and literature, adjusting the preliminary behavior rules of the independent agents, and determining the interaction rules between the independent agents; introducing external environmental impacts and regulatory constraints, and setting a global goal includes: Obtain a structured set of biological interaction information related to independent entities and a normalized literature evidence entry library, compare them with the preliminary behavior rules of the independent entities, dynamically adjust the boundary of the behavior rules of the independent entities, and obtain the behavior rule set of the independent entities; according to the behavior rule set of the independent entities, construct the interaction rules between the independent entities, and clarify the mechanisms such as synergy, antagonism, functional compensation, resource competition, mutual exclusion taboo, and critical dose effect between entities; obtain and quantify the external environmental impact and regulatory constraint parameters, and use them as formula optimization constraints throughout the process. On the basis of forming a coherent mapping with all the behavior rules of the independent entities, construct a global goal.

[0032] In one embodiment, establish a kinetic mapping relationship between the interaction rules of independent entities and the global goal; set the number of iterations and simulation parameters, and in each iteration, each independent entity autonomously adjusts the ratio and actions according to the interaction rules and feedback, and simulates the real - world dynamic cooperation process including: Analyze the interaction rules of independent entities and the global goal, model the impact of the behavior changes and ratio adjustments of each independent entity on the change of the objective function as a causal kinetic mapping; use differential equations to clarify the response relationship between behavior, state, and goal, and form a set of kinetic rules; configure the simulation parameters and initialize the simulation environment. In each round of simulation iteration, all independent entities adjust their own ratios, attributes, and behavior states according to the kinetic mapping model, and calculate the impact of the adjustment results of each independent entity on the global goal in real - time; each independent entity iterates according to its own state and external feedback information, and continuously evolves the dynamic coupling process between the behavior of the independent entity and the global goal; save the historical state, formula combination, and goal results of each iteration for the screening and analysis of the final candidate formula.

[0033] In the process of forming a candidate formula for the plant composite composition, the independent entity has attribute parameters such as functional activity, toxicological characteristics, process adaptability, cost weight, etc. These can be dynamically updated, such as the activity effect being regulated by external feedback of the entity. The preliminary behavior rules are set based on professional literature, databases, and industry experience, including dose - response characteristics, maximum / minimum safe doses, recommended ratio ranges, etc.

[0034] Determine the interaction relationship model between independent entities: cooperation, functional synergism; antagonism, mutual cancellation or negative impact; functional compensation, resource competition, mutual exclusion taboos, dose critical effects, etc. Introduce external parameters (such as temperature, pH, processing intensity, etc.) to act on the behavior rules of the entities. Regulatory parameters, such as maximum allowable dose, list of certified raw materials, and incompatibility taboos, are strictly encoded into the mechanism to determine the upper limit and constraints of the entity behavior. In each round of iteration, each entity self-adjusts its decision variables (such as ratio adjustment, behavior change) according to the latest interaction rules, constraints, and feedback. The independent entities adopt strategies such as game theory, optimization, or genetic evolution to evolve the global objective function towards the optimal (or multi-objective equilibrium) direction. At the same time, record the formula state, global objective value, optimization path, and interaction information between entities in each round. Store all simulation trajectories and formulas that meet the constraint conditions in the database; combine methods such as trajectory analysis and hierarchical clustering to identify typical representative candidate formulas for subsequent multi-objective optimization and final screening.

[0035] In one embodiment, construct a multi-objective optimization framework including effectiveness objectives, stability objectives, safety objectives, cost objectives, and sustainability objectives, and configure the constraint conditions including: According to the determined multi-objective system, configure quantifiable mathematical indicators for each objective (including effectiveness objectives, stability objectives, safety objectives, cost objectives, and sustainability objectives); perform unit normalization on each objective indicator, and assign weights to different objectives to form a weighted objective function; construct all constraint conditions, and solidify the constraint conditions into the input of the optimization algorithm in the form of operators.

[0036] Specifically, embed the normalized multi-objective indicators, weighted objective function, and all constraint operators as inputs into mainstream multi-objective optimization algorithms (such as NSGA-II, MOEA / D, Pareto optimization, etc.); the optimization algorithm filters invalid / infeasible solutions in real time with operators at every key step such as population generation, iteration, and local search to achieve scientific and efficient global optimization and feasible region screening. The multi-objective optimization framework can systematically balance and optimize functions, risks, costs, and sustainability, not only generating theoretically optimal formulas but also ensuring their practical implementation and usability; the flexible adjustable objective system and constraint set can meet the actual industry's rapid change requirements such as new regulations, new raw materials, and new processes.

[0037] In one embodiment, based on the multi-objective optimization algorithm and random forest algorithm, and combined with the multi-objective optimization framework and constraint conditions, rank and screen the candidate formulas of the plant composite composition to obtain a plant composite composition formula set including: Convert each candidate formula of the plant composite composition into a characteristic parameter vector corresponding to a multi-objective index system, and standardize and unify each numerical value; at the same time, filter out all candidate formulas of the plant composite composition that violate the constraint conditions; based on the multi-objective optimization algorithm, comprehensively consider each objective of the multi-objective optimization framework, conduct global optimization and sorting on the candidate formulas of the plant composite composition, and extract a subset of formulas in the optimal trade-off state on the multi-objective solution space; use the random forest model, with the multi-dimensional characteristic parameters of the candidate formulas of the plant composite composition as the input and the composite performance index measured by experiments as the output, to conduct supervised learning training; use the trained random forest model to predict and evaluate the performance of each formula in the formula subset, and obtain the comprehensive performance prediction score corresponding to each formula subset; sort in descending order according to the comprehensive performance prediction score, and output the plant composite composition formula set with the top comprehensive performance ranking and meeting the multi-objective constraint conditions.

[0038] In one embodiment, based on the multi-objective optimization algorithm, comprehensively consider each objective of the multi-objective optimization framework, and conduct global optimization and sorting on the candidate formulas of the plant composite composition, and extracting a subset of formulas in the optimal trade-off state on the multi-objective solution space includes: Select a multi-objective optimization algorithm and initialize the population; run the main loop of the multi-objective optimization algorithm, and use population evolution to make the candidate formulas continuously evolve in the multi-objective solution space; after several iterations, obtain the multi-objective solution set of the multi-objective optimization of the formulas; according to the Pareto optimality principle, identify and extract the candidate formulas of the plant composite composition with the optimal trade-off state of each objective in the formula multi-objective solution set space and that cannot be comprehensively dominated by other solutions, and form a subset of the plant composite composition formulas in the optimal trade-off state.

[0039] In the present invention, by normalizing the candidate formulas of the plant composite composition into an operable characteristic parameter vector, standardize the comparison of various multi-objective performance indicators, strictly combine the pre-set constraint conditions, and automatically eliminate invalid formulas that do not meet the basic requirements. In the multi-objective optimization stage, with the help of mainstream evolutionary algorithms such as NSGA-II, the optimal trade-off can be obtained among different optimization objectives (such as effectiveness, cost, safety, stability, sustainability), and a subset of excellent formulas on the Pareto optimal front can be screened out. Thereafter, with the help of machine learning methods such as random forest, train a prediction model based on the experimental data of previous times, score and sort the performance of the multi-objective subset formulas, so as to further select a high-potential formula set with outstanding comprehensive performance from among many balanced solutions, realize efficient global optimization and intelligent screening driven by data, and effectively improve the scientificity and practical transformation value of formula research and development.

[0040] According to another embodiment of the present invention, there is also provided a plant composite composition, including: 1 - 1.8 parts of licorice root; 0.4 - 1.2 parts of platycodon root; 0.4 - 0.6 parts of polygonatum odoratum; 0.2 - 1.0 parts of phyllanthus emblica; 0.8 - 1.0 parts of dried tangerine peel.

[0041] To facilitate the understanding of the above technical solutions of the present invention, the specific effects and test results of the formula set of the plant composite composition after the formula optimization of the present invention are analyzed as follows: Table 1 Marked components of raw materials

[0042] Analysis of the results of the single - factor test for the formula optimization of the plant composite beverage is as Figure 2 shown. Licorice root contains glycyrrhizin and glycyrrhizic acid. The sweetness of glycyrrhizin is 50 times that of sucrose, but it does not bring the sticky feeling of sugar, giving the plant beverage a natural sweet taste and reducing the use of extra sugar. At the same time, licorice root can balance the sour and astringent taste of phyllanthus emblica and the slightly bitter irritation of dried tangerine peel, making the taste of the beverage smoother and not overly stimulating. As Figure 2 can be seen, when the addition amount of licorice root is 1 - 1.8 g, the sensory score of the plant composite beverage shows a trend of first positive correlation and then negative correlation. When the addition amount of licorice root is 1 - 1.6 g, a small amount of licorice results in poor sweetness of the plant composite beverage and fails to better cover the irritating taste of other plants. When the addition amount of licorice root is 1.6 - 1.8 g, the excessive addition of licorice makes the sweetness of the plant beverage too strong and the color of the beverage too dark. Considering comprehensively, the optimal addition amount of licorice root is 1.6 g, and the sensory score reaches the maximum of 73 points at this time.

[0043] As Figure 3 shown, the influence of the addition amount of platycodon root on the sensory score. Platycodon root itself has a slight bitter and pungent taste, but after being paired with licorice root, the bitterness will be neutralized, leaving a slight aftertaste of sweetness and making the taste more layered. The pungent characteristic of platycodon root gives the beverage a cool feeling similar to mint, making the throat feel soothed and refreshed and reducing the sticky feeling. And an appropriate amount of platycodon root can reconcile the sour and astringent taste of phyllanthus emblica, making the overall aroma milder and more natural. As Figure 3It can be seen that the addition amount of Platycodon grandiflorum in the range of 0.4 - 1.2 g affects the sensory score of the plant compound beverage, showing a trend of first increasing and then decreasing. When the addition amount of Platycodon grandiflorum is 0.8 g, the sensory score reaches the maximum of 72.2 points. When the addition amount of Platycodon grandiflorum is 0.4 - 0.8 g, the sensory score gradually increases. This may be because Platycodon grandiflorum can inhibit the overly sweet taste of Glycyrrhiza glabra. If the amount is too small, the beverage may be too sweet, affecting the taste coordination. At the same time, the herbal fragrance of Platycodon grandiflorum helps to reconcile the sourness of Phyllanthus emblica and the bitterness of Exocarpium Citri Grandis. Too little dosage may lead to an unclear aroma layer, and the sourness of Phyllanthus emblica may be more prominent, thus affecting the overall taste of the plant compound beverage. When the addition amount of Platycodon grandiflorum is 0.8 - 1.2 g, the sensory score gradually decreases because if the slightly pungent characteristic of Platycodon grandiflorum is excessive, it may cause the taste of the beverage to be too stimulating and make the beverage too bitter, thus affecting the sensory score of the plant compound beverage. Considering comprehensively, the optimal addition amount of Platycodon grandiflorum is 0.8 g.

[0044] As Figure 4 shown, the influence of the addition amount of Phyllanthus emblica on the sensory score. Phyllanthus emblica has a strong sour taste, similar to lemon or greengage, and at the same time has a bit of astringency, which comes from its rich tannin components. An appropriate amount of Phyllanthus emblica can enhance the fruit acid flavor of the beverage, making the taste more refreshing, and at the same time can balance the sweetness of Glycyrrhiza glabra and the slight bitterness of Platycodon grandiflorum. From Figure 4 it can be seen that during the process of the addition amount of Phyllanthus emblica in the plant compound beverage increasing from 0.2 g to 1.0 g, the sensory score shows a trend of first increasing and then decreasing. When the addition amount of Phyllanthus emblica is 0.2 - 0.6 g, the sensory score gradually increases because if the addition amount of Phyllanthus emblica is too small, the plant compound beverage may tend to have a herbal taste, reducing the acceptance. As the addition amount of Phyllanthus emblica increases, the sourness neutralizes the sweetness, making the taste more refreshing. When the addition amount of Phyllanthus emblica is 0.6 g, the sensory score reaches the maximum. The sensory score decreases as the addition amount of Phyllanthus emblica is in the range of 0.6 - 1.0 g. It may be because excessive Phyllanthus emblica will cause the plant compound beverage to be overly sour and difficult to accept. High-concentration tannins may react with proteins or other components, making the color of the beverage darker and the transparency lower, thus resulting in a reduction in the overall quality of the plant compound beverage. Considering comprehensively, the addition amount of Phyllanthus emblica is determined to be 0.6 g.

[0045] As Figure 5As shown in the figure, the effect of the addition amount of Polygonatum odoratum on the sensory score. Polygonatum odoratum itself has a relatively sweet, mild taste and a relatively smooth texture. After being added to the beverage, it can make the taste smooth and warm. It also makes the plant compound beverage more gentle and not overly stimulating. Ingredients such as Platycodon grandiflorum and Exocarpium Citri Grandis may bring a certain bitter taste. The addition of Polygonatum odoratum can more effectively neutralize these tastes, making the taste of the beverage more balanced and reducing the prominence of bitterness and astringency. At the same time, Polygonatum odoratum has a faint sweet fragrance, which combines with the sweetness of Glycyrrhiza uralensis to make the overall taste of the beverage milder, reduce the overly strong Chinese medicine taste, and can also balance the aroma, especially being able to cover up the sour and astringent smell of Phyllanthus emblica. From Figure 5 It can be seen that when the addition amount of Polygonatum odoratum is 0.4 - 0.6 g, the sensory score of the plant compound beverage shows an upward trend, and the sensory score reaches the maximum when the addition amount of Polygonatum odoratum is 0.6 g, indicating that the color, shape, taste, and flavor of the jujube, ginger, and Glycyrrhiza uralensis compound beverage are relatively good at this time. When the addition amount of Polygonatum odoratum is 0.6 - 0.8 g, the sensory score of the plant compound beverage shows a downward trend. It may be because the natural sweetness of Polygonatum odoratum is relatively obvious, and excessive use may make the plant compound beverage taste overly sweet and greasy, losing the refreshing effect and affecting the overall taste. Considering comprehensively, the addition amount of Polygonatum odoratum is determined to be 0.6 g.

[0046] As Figure 6 shown in the figure, the effect of the addition amount of Exocarpium Citri Grandis on the sensory score. Exocarpium Citri Grandis itself has a certain bitter taste, and this slightly bitter taste helps to balance with the sweetness of Glycyrrhiza uralensis, avoiding being overly sweet and greasy. And it can enhance the layer of the beverage, making the taste more rich and avoiding monotony. The peel aroma of Exocarpium Citri Grandis has a unique freshness, which can add a natural citrus aroma to the beverage. This aroma can effectively cover up the strong medicinal taste that other medicinal materials may bring, making the overall aroma more pleasant. Exocarpium Citri Grandis usually does not have a significant impact on the color of the beverage when boiled. It mainly releases aroma and a slight bitterness, and its color is relatively light, so it is less likely to change the color of the drink. From Figure 6 It can be seen that when the addition amount of Exocarpium Citri Grandis is 0.8 - 1.0 g, the sensory score of the plant compound beverage shows an upward trend. It may be because the addition amount of Exocarpium Citri Grandis is too small, resulting in the beverage tasting too sweet. The sensory score reaches the maximum value when the addition amount of Exocarpium Citri Grandis is 1.0 g, indicating that the overall taste of the jujube, ginger, and Glycyrrhiza uralensis compound beverage is relatively good at this time. When the addition amount of Exocarpium Citri Grandis is 1.0 - 1.2 g, the sensory score of the plant compound beverage shows a downward trend. It may be because the bitter components of Exocarpium Citri Grandis are relatively significant, and excessive use may cause the taste of the beverage to become overly bitter, affecting the taste and making it difficult to drink. Considering comprehensively, the addition amount of Exocarpium Citri Grandis is determined to be 1.0 g.

[0047] The results of the orthogonal experiment for optimizing the formula of the plant compound beverage are shown in Tables 2 - 3.

[0048] Table 2 Orthogonal experiment design scheme and results of the plant compound beverage

[0049] Table 3 Analysis of Variance

[0050] In Table 2 and Table 3, K represents the total sum, k represents the mean value, and R represents the range; ** indicates extremely significant difference (P<0.001); * indicates significant difference (P<0.05).

[0051] As can be seen from Table 2, by analyzing the differences in R values, it can be seen that there is a significant order among the factors affecting the optimization of the formula of the compound plant beverage, and their primary and secondary relationships are A>B>C>D, that is, the factors affecting the optimization of the formula of the compound plant beverage are the addition amount of licorice, the addition amount of platycodon grandiflorum, the addition amount of phyllanthus emblica, and the addition amount of polygonatum odoratum in turn, and the optimal combination is A2B2C2D1. The formula of the compound plant beverage optimized through the orthogonal experiment is the addition amount of licorice 1.6g, the addition amount of platycodon grandiflorum 0.8g, the addition amount of phyllanthus emblica 0.6g, the addition amount of polygonatum odoratum 0.5g, and the addition amount of citrus reticulata blanco cv. Tomentosa 1.0g. It can be seen from the analysis of variance in Table 3 that the addition amounts of licorice and platycodon grandiflorum have extremely significant effects on the sensory score of the compound plant beverage (P<0.001); the addition amount of phyllanthus emblica has a significant effect on the sensory score of the compound plant beverage (P<0.05); the addition amount of polygonatum odoratum has no significant effect on the sensory score of the compound plant beverage (P>0.05).

[0052] Results of the efficacy verification of the compound plant beverage: (1) Free radicals are harmful substances generated during the oxidation reaction in the body. Their excessive generation will cause cell damage and accelerate individual aging. In the research on antioxidant active substances for scavenging free radicals in the body, it is still a hot field at present. DPPH free radical is a stable molecule and is often used as a tool for measuring the antioxidant capacity of substances, and is widely used in the antioxidant detection of foods, drugs and plants. When the concentration of the compound plant beverage is between 1 and 5 mg / mL, with the continuous increase of the concentration, the scavenging rate of the compound plant beverage on DPPH free radicals gradually increases. When the concentration reaches 5 mg / mL, the scavenging rate reaches the maximum, which is 93.80%. It shows that the compound plant beverage has a certain scavenging effect on DPPH free radicals.

[0053] (2) ABTS free radical is a commonly used antioxidant and can be used to measure the free radical level. When ABTS reacts with free radicals, a reactive intermediate with chromogenic properties will be generated. This intermediate can be detected by a chromatograph, so as to be used to evaluate the scavenging ability of antioxidants on oxidative free radicals. With the increase of the concentration of the compound plant beverage, the scavenging rate on ABTS free radicals is also increasing. When the concentration reaches 5 mg / mL, the scavenging rate of the compound plant beverage reaches the maximum, which is 95.28%. It shows that the compound plant beverage has a certain scavenging effect on ABTS free radicals.

[0054] (3)In the mouse ear swelling test, compared with the negative control group, the ear swelling rate of the mice in the test group with a dose of 0.375 g / kg.bw was significantly reduced (P<0.05). In the experiment of establishing a model of chronic pharyngitis in mice, the test substance had a certain inhibitory effect on the levels of IL-6, IL-1β, and TNF-a in the serum of mice. In the cotton ball implantation test in rats, compared with the positive control group and the negative control group, the net amount of granuloma in the 0.375 g / kg.bw dose group was significantly reduced (P<0.05), and it had a certain inhibitory effect on the levels of IL-6, IL-1β, and TNF-a in the serum of rats.

[0055] According to the evaluation criteria, it can be considered that the beverage had a positive result in the pharyngeal-clearing function test of animal experiments.

[0056] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for optimizing the formula of a plant composite composition, characterized in that, Including: S1. Preprocess multi-domain data of the plant composite composition to obtain multi-source standardized plant feature data; S2. Model each plant raw material component in the multi-source standardized plant feature data as interacting multi-agents; set interaction rules and a global goal, and form a candidate formula for the plant composite composition through a dynamic simulation collaboration process; S3. Construct a multi-objective optimization framework including effectiveness goals, stability goals, safety goals, cost goals, and sustainability goals, and configure constraint conditions; S4. Based on the multi-objective optimization algorithm and the random forest algorithm, and combined with the multi-objective optimization framework and constraint conditions, sort and screen the candidate formulas for the plant composite composition to obtain a formula set for the plant composite composition.

2. The method for optimizing the formula of a plant composite composition according to claim 1, wherein, The preprocessing of multi-domain data of the plant composite composition to obtain multi-source standardized plant feature data includes: Collect multi-domain data of the plant composite composition, including chemical composition data, biological activity data, toxicological data, processing data, and storage data of plant raw materials; Unify the format of the multi-domain data of the plant composite composition, remove noise and outliers, fill in missing values, and perform dimension and unit standardization to form multi-source standardized plant feature data.

3. A method for optimizing the formulation of a plant composite composition according to claim 1, characterized in that, The modeling of each plant raw material component in the multi-source standardized plant feature data as interacting multi-agents; setting interaction rules and a global goal, and forming a candidate formula for the plant composite composition through a dynamic simulation collaboration process includes: Model each plant raw material component in the multi-source standardized plant feature data as an independent agent and endow it with independent agent attributes; determine the preliminary behavior rules of the independent agent and initialize the ratio of each plant raw material component; Adjust the preliminary behavior rules of the independent agent by mining experimental data and literature, and determine the interaction rules between independent agents; introduce external environmental impacts and regulatory constraints, and set a global goal; Establish a kinetic mapping relationship between the interaction rules of independent agents and the global goal; set the number of iterations and simulation parameters, and in each iteration, each independent agent autonomously adjusts the ratio and actions according to the interaction rules and feedback to simulate the real dynamic collaboration process; Record the formula of the plant composite composition in each iteration in real time, and comprehensively form a candidate formula for the plant composite composition.

4. A method for optimizing the formulation of a plant composite composition according to claim 3, characterized in that, The determination of the preliminary behavior rules of the independent agent and the initialization of the ratio of each plant raw material component includes: For each set of independent agent attribute parameters, and combined with literature reports and industry common sense, set the preliminary behavior rules of the independent agent; According to the basic goal of the plant composite composition design, use the normalized allocation method to determine the ratio of each plant raw material component.

5. The method for optimizing the formulation of a plant composite composition according to claim 3, wherein The adjustment of the preliminary behavior rules of the independent agent by mining experimental data and literature, and the determination of the interaction rules between independent agents; the introduction of external environmental impacts and regulatory constraints, and the setting of a global goal includes: Obtain a structured set of biological interaction information related to independent agents and a normalized literature evidence entry library, compare them with the preliminary behavior rules of the independent agent, and dynamically adjust the boundary of the behavior rules of the independent agent to obtain a set of independent agent behavior rules; Construct the interaction rules between independent agents according to the set of independent agent behavior rules; Obtain and quantify the external environmental impacts and regulatory constraint parameters, and use them as constraints for formula optimization throughout the process. Based on forming a coherent mapping with the behavior rules of all independent entities, construct a global objective.

6. The method for optimizing the formulation of a plant composite composition according to claim 3, characterized in that, Establish the dynamic mapping relationship between the interaction rules of independent entities and the global objective; set the number of iterations and simulation parameters, and in each iteration, each independent entity autonomously adjusts the ratio and actions according to the interaction rules and feedback, simulating the real dynamic collaboration process including: Analyze the interaction rules of independent entities and the global objective, and model the impact of the behavior changes and ratio adjustments of each independent entity on the change of the objective function as a causal dynamic mapping; Use differential equations to clarify the response relationships among behavior, state, and objective, and form a set of dynamic rules; Configure the simulation parameters and initialize the simulation environment. In each round of simulation iteration, all independent entities adjust their own ratios, attributes, and behavior states according to the dynamic mapping model, and calculate in real time the impact of the adjustment results of each independent entity on the global objective; Each independent entity iterates according to its own state and external feedback information, and continuously evolves the dynamic coupling process between the behavior of independent entities and the global objective; Save the historical states, formula combinations, and objective results of each iteration for the screening and analysis of the final candidate formulas.

7. A method for optimizing the formulation of a plant composite composition according to claim 1, characterized in that Construct a multi-objective optimization framework including effectiveness objective, stability objective, safety objective, cost objective, and sustainability objective, and configure the constraint conditions including: According to the determined multi-objective system, configure mathematical indicators for the effectiveness objective, stability objective, safety objective, cost objective, and sustainability objective; Perform unit normalization on each objective indicator, and assign weights to different objectives to form a weighted objective function; Construct all constraint conditions, and solidify the constraint conditions into the input of the optimization algorithm in the form of operators.

8. A method for optimizing the formulation of a plant composite composition according to claim 1, characterized in that, Based on the multi-objective optimization algorithm and random forest algorithm, and combined with the multi-objective optimization framework and constraint conditions, sort and screen the candidate formulas of the plant composite composition to obtain the plant composite composition formula set including: Convert each candidate formula of the plant composite composition into a feature parameter vector corresponding to the multi-objective index system, and standardize and unify each numerical value; at the same time, filter out all candidate formulas of the plant composite composition that violate the constraint conditions; Based on the multi-objective optimization algorithm, comprehensively consider each objective of the multi-objective optimization framework, perform global optimization and sorting on the candidate formulas of the plant composite composition, and extract a formula subset in the optimal trade-off state in the multi-objective solution space; Use the random forest model, with the multi-dimensional feature parameters of the candidate formulas of the plant composite composition as the input and the composite performance indicators measured by experiments as the output, to perform supervised learning training; use the trained random forest model to predict and evaluate the performance of each formula in the formula subset, and obtain the comprehensive performance prediction scores corresponding to each formula subset; Arrange in descending order according to the comprehensive performance prediction scores, and output the plant composite composition formula set with the top comprehensive performance ranking and meeting the multi-objective constraint conditions.

9. A method for optimizing the formulation of a plant composite composition according to claim 8, characterized in that, Based on the multi-objective optimization algorithm, comprehensively considering each objective of the multi-objective optimization framework, globally optimize and rank the candidate formulations of the plant composite composition, and extract a subset of formulations in the optimal trade-off state on the multi-objective solution space, including: Select a multi-objective optimization algorithm and initialize the population; run the main loop of the multi-objective optimization algorithm, and use population evolution to continuously evolve the candidate formulations in the multi-objective solution space; After several iterations, obtain the multi-objective solution set of the formulations optimized by the multi-objective optimization; According to the Pareto optimality principle, identify and extract the candidate formulations of the plant composite composition with the optimal trade-off state of each objective in the multi-objective solution set space of the formulations and that cannot be comprehensively dominated by other solutions, to form a subset of formulations in the optimal trade-off state.

10. A plant composite composition, the formulation of the plant composite composition is optimized by using the plant composite composition formulation optimization method described in any one of claims 1-9, characterized in that, The plant composite composition comprises the following components in parts by mass: 1 - 1.8 parts of licorice; 0.4 - 1.2 parts of platycodon root; 0.4 - 0.6 parts of polygonatum odoratum; 0.2 - 1.0 parts of phyllanthus emblica; 0.8 - 1.0 parts of exocarpium citri grandis.

Images