An artificial intelligence-based tree species multi-element mixed carbon sequestration optimization screening method

Abstract

The present application relates to the technical field of tree species optimization screening, and discloses a tree species multi-element mixed carbon sequestration optimization screening method based on artificial intelligence, which comprises the following steps: constructing the single growth characteristics and mixed growth characteristics of tree species according to the pretreated multi-source growth data; predicting the carbon sequestration amount of different tree species combinations in a given growth environment by using a carbon sequestration amount prediction model; calculating the joint growth adaptability and functional complementarity of different tree species combinations, comprehensively scoring the tree species combinations, screening a candidate tree species combination set based on the comprehensive score, constructing a multi-objective optimization function of the candidate tree species combinations in the candidate tree species combination set and solving the function to screen the candidate tree species combination with the highest carbon sequestration efficiency. Through multi-source growth data preprocessing and feature construction, tree species combination carbon sequestration amount prediction, joint scoring and screening and multi-objective optimization, the present application realizes the quantitative evaluation of the carbon sequestration potential of different tree species combinations and the selection of the optimal tree species combination configuration.

Description

Technical Field

[0001] This invention relates to the field of tree species optimization and screening, specifically to a strategy optimization method for tree species screening, and particularly to an artificial intelligence-based method for optimizing carbon sequestration through multi-species hybrid intercropping. Background Technology

[0002] With the increasing severity of global climate change and the continued prominence of ecosystem degradation and environmental pollution, enhancing the carbon sequestration capacity of forests, as important carbon sinks, has become a crucial technical goal for ecological restoration, carbon emission reduction, and sustainable forestry management. Rational allocation and optimization of forest structure can not only increase biomass storage and carbon absorption efficiency but also improve soil quality, hydrological conditions, and biodiversity, thus playing a key role in the overall ecosystem function. Mixed forests, due to their high tree species diversity and complementary characteristics in the utilization of resources such as light, water, and nutrients, and the enhanced resistance to disturbance and stability of forest stands through interspecies interactions, are widely considered to be significantly superior to monoculture stands in terms of carbon sequestration efficiency, ecological benefits, and long-term forest stability.

[0003] In contrast, traditional mixed forest design methods rely primarily on human experience, localized observations, or single indicators for planning, such as planting combinations based solely on tree species growth rates or stand density. This approach struggles to fully consider the complex growth interactions between multiple tree species and the spatial heterogeneity of environmental conditions. Under conditions of multiple tree species, diverse environmental factors, and complex terrain, traditional methods fail to systematically and quantitatively analyze the potential carbon sequestration capacity of forest stands. This results in design schemes lacking scientific rigor and operability, often exhibiting problems such as irrational planting structures, low resource utilization efficiency, and underutilization of carbon sequestration potential. Ultimately, this limits the practical application of mixed forests in carbon sequestration management and ecological restoration.

[0004] Existing research, such as patent CN120197779B, proposes a method and system for optimizing mixed forest stands based on tree cluster units. This method obtains information on tree planting sites in surveyed forests, selects multiple tree cluster units, investigates stand indicators, and constructs a tree species database. It further establishes a correlation model between stand indicators and each pair of relative traits to determine the optimal mixing ratio for each pair of traits under the optimal stand indicators. Planting sites are randomly selected in the proposed planting area, and traits are assigned based on the optimal mixing ratio to match tree species for mixed planting. This method can guide the construction of artificial mixed forests in proposed planting areas from scratch, improving forest management efficiency and replicability. However, this method still has significant technical challenges: it relies on a large amount of field survey data, making it difficult to quickly adapt to different terrains, environmental conditions, and tree species combinations; and it lacks optimization analysis at the carbon sequestration level.

[0005] To address this issue, this invention proposes an artificial intelligence-based method for optimizing carbon sequestration in mixed tree species, which helps to improve the carbon sequestration capacity and productivity of mixed plantations and has significant practical and ecological value for promoting the sustainable use of forest resources. Summary of the Invention

[0006] This invention proposes an artificial intelligence-based method for optimizing and screening carbon sequestration in multi-species mixed planting. Given that traditional methods typically rely on experience or single indicators to evaluate tree species growth potential and carbon sequestration capacity, making it difficult to systematically reflect the comprehensive impact of multiple tree species and environmental factors, this invention collects multi-source growth data and constructs individual and mixed growth characteristics to achieve unified quantification of tree species growth potential, carbon sequestration efficiency, environmental adaptability, cost constraints, and complementary, competitive, and stable synergistic relationships among tree species. Step S2 introduces an AI-based carbon sequestration prediction model, integrating environmental embedding representation, individual feature encoding, and mixed feature encoding, enabling high-precision prediction of carbon sequestration by tree species combinations under a given environment. Step S3 constructs a comprehensive scoring function by combining growth adaptability, functional complementarity, and predicted carbon sequestration values, achieving a scientific, comparable, and quantifiable evaluation of tree species combinations. Step S4 constructs a multi-objective optimization function constrained by tree species planting area to achieve the optimal combination configuration that balances carbon sequestration efficiency and plantability.

[0007] To achieve the above objectives, this invention provides an artificial intelligence-based method for optimizing and screening carbon sequestration through multi-species mixed intercropping, comprising the following steps: S1: Collect multi-source growth data of different tree species, and preprocess the multi-source growth data to obtain preprocessed multi-source growth data. Based on the preprocessed multi-source growth data, construct the individual growth characteristics and mixed growth characteristics of the tree species. S2: Based on the individual growth characteristics and mixed growth characteristics of the tree species, the carbon sequestration prediction model is used to predict the carbon sequestration values ​​of different tree species combinations under a given growth environment. S3: Based on the individual growth characteristics and mixed growth characteristics of the tree species, calculate the joint growth adaptability and functional complementarity of different tree species combinations. Combine the predicted carbon sequestration values ​​of the different tree species combinations under a given growth environment, and give a comprehensive score to the tree species combinations. Based on the comprehensive score, select a set of candidate tree species combinations. S4: Using the planting area of ​​tree species as a constraint, construct a multi-objective optimization function for the candidate tree species combination in the candidate tree species combination set. By solving the multi-objective optimization function, optimize and screen to obtain the candidate tree species combination with the highest carbon sequestration efficiency.

[0008] As a further improvement of the present invention: Furthermore, in step S1, multi-source growth data of different tree species are collected, and the multi-source growth data is preprocessed, including: The multi-source growth data includes growth characteristic data of tree species when they reach the estimated growth stage, historical carbon sequestration data, growth environment factor data, and operating cost data; The growth characteristic data include the average tree height, average diameter at breast height (DBH), and average root depth of the tree species. The historical carbon sequestration data includes the change in biomass of tree species per unit time and historical carbon sink monitoring records. The growth environment factor data includes the optimal environmental factor index values ​​for tree species growth of different types of environmental factors and the actual environmental factor index values ​​of the target planting area under different types of environmental factors, wherein the types of environmental factors include light, water and nutrients. The operating cost data includes the cost of purchasing seedlings per unit planting area, the cost of planting labor, and the cost of maintenance and management when the tree species reaches the estimated growth stage. The multi-source growth data is preprocessed, wherein the preprocessing procedure includes: S11: Complete missing data using interpolation or historical mean; S12: Normalize the growth data of different dimensions to obtain preprocessed multi-source growth data.

[0009] Furthermore, based on the preprocessed multi-source growth data, individual growth characteristics and mixed growth characteristics of tree species are constructed, including: S13: Based on the preprocessed multi-source growth data, calculate the individual growth characteristics of the tree species, wherein the individual growth characteristics include growth potential characteristics, carbon sequestration efficiency characteristics, environmental adaptability characteristics, and cost constraint characteristics. S14: Based on the individual growth characteristics of the tree species, introduce the interaction relationship between tree species to construct the mixed growth characteristics between any two tree species. The mixed growth characteristics include growth complementarity characteristics, competition inhibition characteristics, and stability synergy characteristics. Growth complementarity characteristics describe the degree of complementarity between different tree species in spatial structure or resource utilization. Competition inhibition characteristics describe the competition relationship between tree species for light, water, and nutrients. Stability synergy characteristics describe the impact of the mixed structure on community stability and resistance to disturbance.

[0010] Furthermore, the carbon fixation prediction model in step S2 includes a growth environment coding module, a single-unit feature coding module, a mixed growth feature coding module, and a feature fusion and nonlinear mapping module, comprising: The growth environment coding module is used to perform nonlinear mapping on the actual environmental factor index values ​​of the target planting area to form an environmental embedding representation that can characterize the overall growth conditions of the area. The individual feature encoding module is used to encode the growth potential, carbon sequestration efficiency, environmental adaptability and cost constraint characteristics of each tree species in the tree species combination, and to use an attention mechanism to weight and fuse the encoding results of different tree species to highlight the individual features of tree species that contribute more to the current growth environment. The mixed growth feature encoding module is used to encode and aggregate the growth complementarity features, competition inhibition features, and stability synergy features between any two tree species in the tree species combination, so as to characterize the interaction relationship under mixed tree species conditions; The feature fusion and nonlinear mapping module is used to splice and fuse environmental embedding representations, single-unit encoded features and hybrid encoded features, and to achieve adaptive adjustment of different feature information through a multi-layer nonlinear mapping structure based on a gating mechanism, and output the predicted carbon sequestration value of the tree species combination under a given growth environment.

[0011] Furthermore, step S2, which uses a carbon sequestration prediction model to predict the predicted carbon sequestration values ​​of different tree species combinations under a given growth environment, also includes: S21: Divide the tree species into multiple tree species combinations according to the preset range of mixed tree species; S22: Based on the individual growth characteristics and mixed growth characteristics of the tree species, extract the individual growth characteristics of each tree species in the tree species combination and the mixed growth characteristics between any two tree species. S23: The actual environmental factor index values ​​of the target planting area under different types of environmental factors are taken as the given growth environment. The growth environment coding module performs nonlinear mapping on the given growth environment to obtain the environmental embedding representation. Specifically, the given growth environment is represented as follows: ,in The values ​​of the first to third environmental factors in the target planting area, respectively, represent the normalized values ​​after normalization. These environmental factors are light, water, and nutrients, respectively. The given growth environment... The formula for nonlinear mapping is: ; in, Indicates a given growth environment The corresponding context embedding representation, This represents the trainable environment mapping weight matrix. This represents the bias of the trainable environment mapping. Represents the hyperbolic tangent function; S24: The individual feature encoding module encodes the individual growth features of each tree species in the tree species combination, and fuses the encoding results based on the attention mechanism to obtain the individual encoded features of the tree species combination. ; ; in, Indicates tree species combination The single-character coding features, Indicates tree species combination The j-th tree species The characteristics of single-cell growth, , Indicates tree species combination The number of tree species in the middle, Indicates tree species combination The j-th tree species Attention weights Indicates monomer growth characteristics The encoding result, This represents the encoding matrix in a trainable single-unit feature encoding module. This represents the encoding bias in a trainable single-unit feature encoding module. This represents the first activation function, set It is the ReLU activation function. Represents an exponential function with the natural constant as its base; S25: The mixed growth feature encoding module encodes and fuses the mixed growth features in the tree species combination to obtain the mixed encoding features of the tree species combination; Specifically, the formula for generating the hybrid coding features of the tree species combination is as follows: ; in, Indicates tree species combination Hybrid coding features, , Indicates tree species combination Different tree species in the middle, Indicate tree species Mixed growth characteristics between This represents the encoding matrix in a trainable hybrid growth feature encoding module. This represents the encoding bias in a trainable hybrid growth feature encoding module; S26: The feature fusion and nonlinear mapping module concatenates the environment embedding representation, the individual coding features of the tree species combination, and the hybrid coding features of the tree species combination, and outputs the predicted carbon sequestration value of the tree species combination using a multi-layer nonlinear mapping method based on a gating mechanism.

[0012] Specifically, the feature fusion and nonlinear mapping module includes R gated mapping layers, as shown in the figure. Figure 3The gating mapping flowchart shown is as follows, where the calculation process for the r-th gating mapping layer is: Calculate the r-th gate vector: ; in, Let r be the gating vector. This represents the second activation function, set It is the Sigmoid activation function. This represents the gate weight matrix in the r-th gated mapping layer. This represents the gating bias in the r-th gating mapping layer. This represents the nonlinear mapping result output by the (r-1)th gated mapping layer, set... ; Map the nonlinear mapping result output from the previous gating mapping layer. ; in, This represents the nonlinear mapping weight matrix in the r-th gated mapping layer. This represents the nonlinear mapping bias in the r-th gated mapping layer. Represents the result of nonlinear mapping The mapping value; The mapping result is then gated and fused with the gated vector to obtain the nonlinear mapping result of the current gated mapping layer. ; in, Indicate length and A unit vector that is identical and whose vector value is 1. This represents the Hadamard product operator. This represents the nonlinear mapping result output by the r-th gated mapping layer; The nonlinear mapping result output by the Rth gated mapping layer is converted into a carbon fixation prediction value using a fully connected layer.

[0013] Furthermore, step S3 calculates the combined growth adaptability and functional complementarity of different tree species combinations, including: S31: Calculate the contribution weight coefficient of the tree species in the combination of tree species based on the individual growth characteristics and mixed growth characteristics of the tree species. S32: Based on the contribution weight coefficient of the tree species, perform weighted mapping on the individual growth characteristics of the tree species in the tree species combination to obtain the joint growth adaptability of the tree species combination; S33: Based on the contribution weight coefficient of the tree species, calculate the comprehensive contribution weight coefficient of any two tree species in the tree species combination, perform weighted mapping on the mixed growth characteristics between tree species, and obtain the functional complementarity of the tree species combination.

[0014] Furthermore, step S3, which combines the predicted carbon sequestration values ​​of different tree species combinations under a given growth environment to comprehensively score the tree species combinations, also includes: S34: Obtain the predicted carbon sequestration value of the tree species combination under a given growth environment; S35: Based on the combined growth adaptability, functional complementarity, and predicted carbon sequestration of the tree species combination, a comprehensive score is given to the tree species combination using a joint scoring function; S36: Based on the comprehensive score, select the top 10% of tree species combinations with the highest comprehensive scores as candidate tree species combinations, and construct a candidate tree species combination set.

[0015] Furthermore, in step S4, a multi-objective optimization function is constructed for the candidate tree species combinations in the candidate tree species combination set, using the planting area of ​​the tree species as a constraint. The expression of the multi-objective optimization function is as follows: ; ; in, Represents a multi-objective optimization function. Indicates candidate tree species combinations The multi-objective optimization function value, Indicates candidate tree species combinations The predicted value of carbon sequestration, Indicates candidate tree species combinations Overall score This represents the maximum value among the comprehensive scores of all candidate tree species combinations in the candidate tree species combination set. Indicates candidate tree species combinations The d-th tree species The number of plants, Indicates candidate tree species combinations The d-th tree species The planting area of ​​a single tree This indicates the arable area of ​​the target planting region. This indicates the minimum number of tree species to be planted. Indicates candidate tree species combinations The number of tree species, including candidate tree species combinations The number of trees planted for each species is the parameter to be solved in the multi-objective optimization function; This indicates a constraint.

[0016] Compared with existing technologies, this invention proposes an artificial intelligence-based method for optimizing and screening carbon fixation through multi-species mixed intercropping. This technology has the following beneficial effects: First, this invention, based on artificial intelligence, constructs a carbon sequestration prediction model that includes growth environment coding, individual feature coding, mixed growth feature coding, feature fusion, and nonlinear mapping. This model achieves refined modeling and prediction of carbon sequestration potential under mixed tree species conditions. Specifically, the model first uses a growth environment coding module to nonlinearly embed environmental factors such as light, water, and nutrients, effectively characterizing the overall constraints of regional growth conditions. The individual feature coding module, considering tree species growth potential, carbon sequestration efficiency, environmental adaptability, and cost constraints, introduces an attention mechanism to adaptively highlight dominant tree species that contribute significantly to the current environment. The mixed growth feature coding module systematically models the complementary, competitive, and synergistic relationships among tree species, compensating for the shortcomings of traditional models that ignore mixed-species effects. Through a multi-layer nonlinear mapping structure based on a gating mechanism, the model achieves dynamic adjustment and efficient fusion of multi-source features, improving the stability, accuracy, and generalization ability of carbon sequestration prediction under complex ecological conditions.

[0017] Meanwhile, this invention introduces a multi-objective optimization function with tree species planting area as a constraint, unifying the carbon sequestration prediction results, comprehensive scoring results, and actual planting space conditions of tree species combinations into a unified model, thus achieving effective connection from combination evaluation to engineering configuration. Specifically, the multi-objective optimization function uses the carbon sequestration prediction value as the core benefit term, and combines the comprehensive score to normalize and modulate the ecological adaptability and structural rationality of different tree species combinations, highlighting the carbon sequestration efficiency advantage while ensuring the fairness of combination evaluation; furthermore, this invention avoids the problem of optimization results being unimplementable in actual planting by explicitly introducing planting area constraints and minimum planting quantity constraints. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating an artificial intelligence-based method for optimizing and screening carbon fixation through multi-species hybrid intercropping of trees, as provided in an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of a carbon sequestration prediction model provided in an embodiment of the present invention.

[0020] Figure 3 This is a flowchart of a gating mapping process provided in an embodiment of the present invention.

[0021] Figure 4 This is a schematic diagram of row-shaped mixed planting provided in an embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of block-shaped mixed planting provided in an embodiment of the present invention. Detailed Implementation

[0023] The realization of the objectives, functional characteristics, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0024] This invention provides an artificial intelligence-based method for optimizing carbon sequestration in multi-species mixed planting. The executing entity of this artificial intelligence-based method includes, but is not limited to, at least one electronic device configured to execute the method provided in this invention, such as a server or a terminal. In other words, the artificial intelligence-based method can be executed by software or hardware installed on a terminal device or a server device, and the software may be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0025] Reference Figure 1 , Figure 2 as well as Figure 3 Embodiment 1 of the present invention is as follows: An artificial intelligence-based method for optimizing and screening carbon sequestration through multi-species mixed intercropping, the method comprising: S1: Collect multi-source growth data of different tree species, and preprocess the multi-source growth data to obtain preprocessed multi-source growth data. Based on the preprocessed multi-source growth data, construct the individual growth characteristics and mixed growth characteristics of the tree species.

[0026] Step S1 involves collecting multi-source growth data from different tree species and preprocessing the multi-source growth data, including: The multi-source growth data includes growth characteristic data of tree species when they reach the estimated growth stage, historical carbon sequestration data, growth environment factor data, and operating cost data; In one embodiment of the present invention, the estimated growth stage refers to a standardized growth period node pre-set for the purpose of unifying the comparison of multi-source growth data for different tree species. The years corresponding to the estimated growth stage are: ,in Five years indicates that the tree species is in the rapid growth stage of young forests, suitable for early carbon sequestration potential assessment; 8-10 years indicates that the tree species has entered a stable growth stage, suitable for mainstream carbon sequestration optimization decisions; 15 years indicates that the tree species is in the middle-aged forest stage, used for long-term carbon sequestration capacity assessment, and preferred settings. For 10 years; The growth characteristic data include the average tree height, average diameter at breast height (DBH), and average root depth of the tree species. Specifically, the average diameter at breast height (DBH) represents the arithmetic mean of the trunk diameters of the tree species measured at a distance of 1.3 meters from the ground. The historical carbon sequestration data includes the change in biomass of tree species per unit time and historical carbon sink monitoring records. Specifically, the method for collecting the biomass change per unit time of the tree species is as follows: ; in, All of these represent empirical coefficients. This represents the average change in diameter at breast height (DBH) of a tree species per unit time. This represents the average change in tree height per unit time. This represents the change in biomass of a tree species per unit time, where represents the empirical scale coefficient used to map geometric magnitudes to biomass magnitudes. This represents the empirical coefficient of diameter at breast height (DBH), reflecting the sensitivity of biomass to changes in DBH. The empirical coefficient for tree height reflects the contribution of biomass to height growth. It can be obtained by performing logarithmic regression on historical biomass, diameter at breast height (DBH), and tree height data of the tree species, or preset according to the tree species growth type. Optionally, the empirical coefficient settings for different tree species growth types are as follows: Tree species growth type: fast-growing broadleaf; empirical coefficient of scale: 0.2; empirical coefficient of diameter at breast height: 2.4; empirical coefficient of tree height: 0.9; Tree species growth type: mesophyllary broadleaf; empirical coefficient of scale: 0.3; empirical coefficient of diameter at breast height: 2.2; empirical coefficient of tree height: 0.8; Tree species growth type: coniferous; empirical coefficient of scale: 0.35; empirical coefficient of diameter at breast height: 2.1; empirical coefficient of tree height: 0.6; The growth environment factor data includes the optimal environmental factor index values ​​for tree species growth of different types of environmental factors and the actual environmental factor index values ​​of the target planting area under different types of environmental factors, wherein the types of environmental factors include light, water and nutrients. Specifically, the nutrients include nitrogen, phosphorus and potassium, wherein nitrogen affects leaf formation, photosynthetic efficiency and growth rate, phosphorus affects root development, energy metabolism and stress resistance, and potassium affects water regulation, enzyme activity and stress resistance stability. As an embodiment of the present invention, the optimal environmental factor index value for tree species growth of the different types of environmental factors is obtained by: statistically analyzing the average environmental factor value of tree species in high-growth areas based on historical sample plot data, and using it as the corresponding optimal environmental factor index value. The method for obtaining the actual environmental factor index values ​​of the target planting area under different types of environmental factors is as follows: the actual environmental factor index values ​​of the target planting area under different types of environmental factors are measured using a soil sensor. The operating cost data includes the cost of purchasing seedlings per unit planting area, the cost of planting labor, and the cost of maintenance and management when the tree species reaches the estimated growth stage. The multi-source growth data is preprocessed, wherein the preprocessing procedure includes: S11: Complete missing data using interpolation or historical mean; S12: Normalize the growth data of different dimensions to obtain preprocessed multi-source growth data.

[0027] In one embodiment of the present invention, the normalization method for growth characteristic data, carbon sequestration history data, growth environment factor data, and operating cost data is the ratio between the actual value and the maximum value among all tree species. Taking the average tree height H of the tree species as an example, the normalization formula for the average tree height H is as follows: ,in This represents the maximum average height of all tree species.

[0028] Based on the preprocessed multi-source growth data, individual growth characteristics and mixed growth characteristics of tree species are constructed, including: S13: Based on the preprocessed multi-source growth data, calculate the individual growth characteristics of the tree species, wherein the individual growth characteristics include growth potential characteristics, carbon sequestration efficiency characteristics, environmental adaptability characteristics, and cost constraint characteristics. Specifically, the calculation method for the individual growth characteristics of the tree species is as follows: ; ; ; ; ; in, Indicates the first Individual growth characteristics of tree species, where N represents the number of tree species. They represent the number respectively. The growth potential, carbon sequestration efficiency, environmental adaptability, and cost constraints of different tree species were considered. Indicates the number of digits after normalization. The average diameter at breast height (DBH) of the tree species planted Indicates the number of digits after normalization. Average height of the planted trees Indicates the growth ratio coefficient, set It is 0.4. Indicates the number of digits after normalization. Change in biomass per unit time for planted tree species This represents the biomass-carbon conversion factor (which can be set to 0.4). Indicates the number of digits after normalization. The optimal environmental factor index value for planting trees under the k-th environmental factor. This represents the actual environmental factor index value of the target planting area for the kth environmental factor after normalization treatment, where The first to third environmental factors are, in order, light, water, and nutrients. These represent the normalized values ​​respectively. The costs of purchasing seedlings for planting trees, the labor costs of planting, and the costs of maintenance and management; S14: Based on the individual growth characteristics of the tree species, introduce the interaction relationship between tree species to construct the mixed growth characteristics between any two tree species. The mixed growth characteristics include growth complementarity characteristics, competition inhibition characteristics, and stability synergy characteristics. Growth complementarity characteristics describe the degree of complementarity between different tree species in spatial structure or resource utilization. Competition inhibition characteristics describe the competition relationship between tree species for light, water, and nutrients. Stability synergy characteristics describe the impact of the mixed structure on community stability and resistance to disturbance.

[0029] Specifically, the calculation method for the mixed growth characteristics between any two tree species is as follows: ; ; ; ; in, Indicates the first Mixed growth characteristics between tree species n and the nth tree species. They represent the number respectively. The complementary growth characteristics, competitive inhibition characteristics, and stable synergistic characteristics between tree species n and the nth tree species. Indicates the number of digits after normalization. Average root depth of the tree species This represents the average root depth of the nth tree species after normalization. Indicates the number of digits after normalization. A sequence of optimal environmental factor index values ​​for tree species. This represents the sequence of optimal environmental factor index values ​​for the nth tree species after normalization. These represent the normalized values ​​respectively. The optimal environmental factor index values ​​for planting trees in the first to third environmental factors. These represent the normalized values ​​respectively. The optimal environmental factor index values ​​for planting trees in the first to third environmental factors. Indicates the number of digits after normalization. Standard deviation of historical carbon sequestration monitoring records for tree species planted This represents the standard deviation of historical carbon sink monitoring records for the nth tree species after normalization.

[0030] It should be noted that this invention structurates and quantifies the individual growth capacity of tree species and the interaction between tree species, effectively solving the problem of relying solely on experience or single indicators for evaluation in traditional tree species configuration. At the individual tree species level, this invention decouples and models growth potential, carbon sequestration efficiency, environmental adaptability, and cost constraints, enabling a unified characterization and comparison of differences among different tree species in terms of growth potential, carbon sink contribution, environmental compatibility, and operational feasibility, thus improving the objectivity and calculability of basic tree species evaluation. At the mixed tree species level, this invention systematically describes the complementary, competitive, and stable synergistic relationships among tree species by introducing differences in root system hierarchy, similarity of environmental demand vectors, and historical carbon sink fluctuation characteristics. This allows for explicit modeling of the impact of mixed structures on resource utilization efficiency and community stability, providing a highly discriminative and interpretable feature foundation for subsequent joint growth adaptability assessment, carbon sequestration prediction, and optimal selection of tree species combinations, significantly improving the scientific rigor and engineering applicability of multi-species mixed carbon sequestration optimization methods.

[0031] S2: Based on the individual growth characteristics and mixed growth characteristics of the tree species, the carbon sequestration prediction model is used to predict the carbon sequestration values ​​of different tree species combinations under a given growth environment.

[0032] For reference Figure 2 The diagram shows the structure of a carbon sequestration prediction model. This model includes a growth environment coding module, a single-unit feature coding module, a mixed growth feature coding module, and a feature fusion and nonlinear mapping module. The growth environment coding module is used to perform nonlinear mapping on the actual environmental factor index values ​​of the target planting area to form an environmental embedding representation that can characterize the overall growth conditions of the area. The individual feature encoding module is used to encode the growth potential, carbon sequestration efficiency, environmental adaptability and cost constraint characteristics of each tree species in the tree species combination, and to use an attention mechanism to weight and fuse the encoding results of different tree species to highlight the individual features of tree species that contribute more to the current growth environment. The mixed growth feature encoding module is used to encode and aggregate the growth complementarity features, competition inhibition features, and stability synergy features between any two tree species in the tree species combination, so as to characterize the interaction relationship under mixed tree species conditions; The feature fusion and nonlinear mapping module is used to splice and fuse environmental embedding representations, single-unit encoded features and hybrid encoded features, and to achieve adaptive adjustment of different feature information through a multi-layer nonlinear mapping structure based on a gating mechanism, and output the predicted carbon sequestration value of the tree species combination under a given growth environment.

[0033] Step S2, which uses a carbon sequestration prediction model to predict the carbon sequestration values ​​of different tree species combinations under a given growth environment, also includes: S21: Divide the tree species into multiple tree species combinations according to the preset range of mixed tree species; Specifically, if the preset range of the number of mixed tree species is 3-8, then the divided tree species combinations must contain at least 3 different tree species and at most 8 different tree species. S22: Based on the individual growth characteristics and mixed growth characteristics of the tree species, extract the individual growth characteristics of each tree species in the tree species combination and the mixed growth characteristics between any two tree species. S23: The actual environmental factor index values ​​of the target planting area under different types of environmental factors are taken as the given growth environment. The growth environment coding module performs nonlinear mapping on the given growth environment to obtain the environmental embedding representation. Specifically, the given growth environment is represented as follows: ,in These represent the normalized values ​​of the target planting area for the first to third environmental factors, respectively. The given growth environment... The formula for nonlinear mapping is: ; in, Indicates a given growth environment The corresponding context embedding representation, This represents the trainable environment mapping weight matrix. This represents the bias of the trainable environment mapping. Represents the hyperbolic tangent function; S24: The individual feature encoding module encodes the individual growth features of each tree species in the tree species combination, and fuses the encoding results based on the attention mechanism to obtain the individual encoded features of the tree species combination. Specifically, the formula for generating the individual encoding features of the tree species combination is as follows: ; ; in, Indicates tree species combination The single-character coding features, Indicates tree species combination The j-th tree species The characteristics of single-cell growth, , Indicates tree species combination The number of tree species in the middle, Indicates tree species combination The j-th tree species Attention weights Indicates monomer growth characteristics The encoding result, This represents the encoding matrix in a trainable single-unit feature encoding module. This represents the encoding bias in a trainable single-unit feature encoding module. This represents the first activation function, set It is the ReLU activation function. Represents an exponential function with the natural constant as its base; S25: The mixed growth feature encoding module encodes and fuses the mixed growth features in the tree species combination to obtain the mixed encoding features of the tree species combination; Specifically, the formula for generating the hybrid coding features of the tree species combination is as follows: ; in, Indicates tree species combination Hybrid coding features, , Indicates tree species combination Different tree species in the middle, Indicate tree species Mixed growth characteristics between This represents the encoding matrix in a trainable hybrid growth feature encoding module. This represents the encoding bias in a trainable hybrid growth feature encoding module; S26: The feature fusion and nonlinear mapping module concatenates the environment embedding representation, the individual coding features of the tree species combination, and the hybrid coding features of the tree species combination, and outputs the predicted carbon sequestration value of the tree species combination using a multi-layer nonlinear mapping method based on a gating mechanism.

[0034] Specifically, the feature fusion and nonlinear mapping module includes R gated mapping layers, as shown in the figure. Figure 3 The gating mapping flowchart shown is as follows, where the calculation process for the r-th gating mapping layer is: Calculate the r-th gate vector: ; in, Let r be the gating vector. This represents the second activation function, set It is the Sigmoid activation function. This represents the gate weight matrix in the r-th gated mapping layer. This represents the gating bias in the r-th gating mapping layer. This represents the nonlinear mapping result output by the (r-1)th gated mapping layer, set... ; Map the nonlinear mapping result output from the previous gating mapping layer: ; in, This represents the nonlinear mapping weight matrix in the r-th gated mapping layer. This represents the nonlinear mapping bias in the r-th gated mapping layer. Represents the result of nonlinear mapping The mapping value; The mapping result is then gated and fused with the gated vector to obtain the nonlinear mapping result of the current gated mapping layer. ; in, Indicate length and A unit vector that is identical and whose vector value is 1. This represents the Hadamard product operator. This represents the nonlinear mapping result output by the r-th gated mapping layer; The nonlinear mapping result output by the Rth gated mapping layer is converted into a carbon fixation prediction value using a fully connected layer.

[0035] As an embodiment of the present invention, by collecting the given growth environment, individual growth characteristics, mixed growth characteristics, and actual carbon sequestration of tree species combinations under multiple mixed planting scenarios, wherein the prediction time period is 3 years, a training loss function is constructed with the objective of minimizing the difference between the predicted carbon sequestration value output by the carbon sequestration prediction model and the actual carbon sequestration of the prediction time period, and the trainable parameters in the carbon sequestration prediction model are trained and optimized. The training optimization method is gradient descent algorithm or Adam optimizer.

[0036] It should be noted that the carbon sequestration prediction value output by the carbon sequestration prediction model is the carbon sequestration prediction value generated after the length of the prediction time period following the mixed planting of tree species combinations.

[0037] S3: Based on the individual growth characteristics and mixed growth characteristics of the tree species, calculate the joint growth adaptability and functional complementarity of different tree species combinations. Combine the predicted carbon sequestration values ​​of the different tree species combinations under a given growth environment, and give a comprehensive score to the tree species combinations. Based on the comprehensive score, select a set of candidate tree species combinations.

[0038] The S3 step calculates the combined growth adaptability and functional complementarity of different tree species combinations, including: S31: Calculate the contribution weight coefficient of the tree species in the combination of tree species based on the individual growth characteristics and mixed growth characteristics of the tree species. As an embodiment of the present invention, the calculation process for the contribution weight coefficient of the tree species in the tree species combination is as follows: S311: Calculate the individual comprehensive adaptability index of each tree species in the tree species combination; wherein the tree species combination The j-th tree species Individual comprehensive adaptability index for: ; in, The tree species combinations are represented in sequence. The growth potential, carbon sequestration efficiency, environmental adaptability, and cost constraints of the j-th tree species are analyzed. All represent the adaptive weighting coefficients, set In order to be ; S312: Convert the individual comprehensive adaptability index of each tree species in the tree species combination into the contribution weight coefficient of each tree species; wherein the tree species combination The j-th tree species Contribution weighting coefficient for: ; in, Indicates tree species combination The number of tree species in the area; S32: Based on the contribution weight coefficient of the tree species, perform weighted mapping on the individual growth characteristics of the tree species in the tree species combination to obtain the joint growth adaptability of the tree species combination; Specifically, the formula for calculating the joint growth adaptability of the tree species combination S is as follows: ; in, This indicates the combined growth adaptability of the tree species combination S. Indicates tree species combination The j-th tree species With the m-th tree species Stability and synergistic characteristics between them; S33: Based on the contribution weight coefficient of the tree species, calculate the comprehensive contribution weight coefficient of any two tree species in the tree species combination, perform weighted mapping on the mixed growth characteristics between tree species, and obtain the functional complementarity of the tree species combination.

[0039] As an embodiment of the present invention, the tree species combination The j-th tree species With the m-th tree species The overall contribution weighting coefficient between them is: ; in, Indicates the combination of tree species The j-th tree species With the m-th tree species The overall contribution weighting coefficient between them; The formula for calculating the functional complementarity of the tree species combination S is: ; in, The tree species combinations are as follows: The j-th tree species With the m-th tree species Complementary growth characteristics and competitive inhibition characteristics between them Indicates the growth control coefficient, set It is 0.8. This indicates the functional complementarity of the tree species combination S.

[0040] It should be noted that this invention constructs a single comprehensive adaptability index that includes growth potential, carbon fixation efficiency, environmental adaptability and cost constraints, and maps it to the contribution weight coefficient of tree species. This allows the role of different tree species in the combination to be adaptively allocated according to their comprehensive advantages, avoiding the evaluation bias caused by the simple equal weighting of the contributions of each tree species in the traditional method. In the calculation of joint growth adaptability, this invention introduces stability synergy features to modulate the individual adaptability index, effectively reflecting the positive effect of mutual promotion among tree species on the overall growth stability under mixed conditions. In the functional complementarity assessment, the complementary growth, competitive inhibition, and stability synergy characteristics among tree species are weighted and integrated by comprehensively contributing weights. This allows tree species pairs with prominent complementary advantages to receive higher weights in the evaluation results, while suppressing the scores of highly competitive combinations. This significantly improves the distinguishability and reliability of different tree species combinations, providing a quantitative decision-making basis that balances ecological benefits and operational feasibility for the optimal screening of multi-species mixed carbon sequestration configurations.

[0041] Step S3, which combines the predicted carbon sequestration values ​​of different tree species combinations under a given growth environment to perform a comprehensive evaluation of the tree species combinations, also includes: S34: Obtain the predicted carbon sequestration value of the tree species combination under a given growth environment; S35: Based on the combined growth adaptability, functional complementarity, and predicted carbon sequestration of the tree species combination, a comprehensive score is given to the tree species combination using a joint scoring function; As an embodiment of the present invention, the functional expression of the joint scoring function is: ; in, This represents the joint scoring function. This represents the overall score of the tree species combination S. This represents the predicted carbon sequestration value of the tree species combination S. This represents the standard deviation of the individual comprehensive adaptability index of each tree species in the tree species combination S. This represents the penalty term for the tree species combination S. Indicates the synergistic amplification factor of growth adaptability. Indicates the synergistic amplification factor of functional complementarity. Indicates the penalty coefficient, set The values ​​are 1.1, 1.2, and 0.8 respectively. It should be noted that this invention upgrades the comprehensive evaluation of tree species combinations from the traditional linear weighted model to a risk-aware nonlinear assessment model by introducing a joint scoring function based on multiplicative synergy and penalty mechanisms. Specifically, this scoring function uses the predicted carbon sequestration value as the core driving indicator and synergistically amplifies it through the combination of growth adaptability and functional complementarity. This can truly reflect the promoting effect of adaptive advantages and functional differentiation on the release of carbon sequestration potential under mixed tree species conditions. At the same time, this invention effectively reduces the risk of structural instability caused by tree species capacity polarization by quantitatively characterizing and exponentially penalizing the differences in adaptability among tree species within the combination. That is, it suppresses the risk of structural imbalance caused by the dominance of a single dominant tree species and improves the reliability of the screening results under long-term management and climate fluctuation conditions.

[0042] S36: Based on the comprehensive score, select the top 10% of tree species combinations with the highest comprehensive scores as candidate tree species combinations, and construct a candidate tree species combination set.

[0043] S4: Using the planting area of ​​tree species as a constraint, construct a multi-objective optimization function for the candidate tree species combination in the candidate tree species combination set. By solving the multi-objective optimization function, optimize and screen to obtain the candidate tree species combination with the highest carbon sequestration efficiency.

[0044] In step S4, a multi-objective optimization function is constructed for the candidate tree species combinations in the candidate tree species combination set, using the planting area of ​​the tree species as a constraint. The expression of the multi-objective optimization function is as follows: ; ; in, Represents a multi-objective optimization function. Indicates candidate tree species combinations The multi-objective optimization function value, Indicates candidate tree species combinations The predicted value of carbon sequestration, Indicates candidate tree species combinations Overall score This represents the maximum value among the comprehensive scores of all candidate tree species combinations in the candidate tree species combination set. Indicates candidate tree species combinations The d-th tree species The number of plants, Indicates candidate tree species combinations The d-th tree species The planting area of ​​a single tree This indicates the arable area of ​​the target planting region. This indicates the minimum number of tree species to be planted (can be set to 10). Indicates candidate tree species combinations The number of tree species, including candidate tree species combinations The number of trees planted for each species is the parameter to be solved in the multi-objective optimization function; This indicates a constraint.

[0045] As an embodiment of the present invention, for each candidate tree species combination, with the goal of maximizing the multi-objective optimization function value, a heuristic algorithm (such as a genetic algorithm or particle swarm optimization algorithm) is used to optimize and solve the multi-objective optimization function to obtain the number of different tree species to be planted. The number of trees to be planted is then substituted into the multi-objective optimization function to obtain the optimized multi-objective optimization function value corresponding to the candidate tree species combination. The candidate tree species combination with the largest optimized multi-objective optimization function value is selected as the candidate tree species combination with the highest carbon sequestration efficiency.

[0046] As another preferred embodiment of the present invention, a differentiated mixed-species configuration is adopted for the selected tree species combinations according to the terrain conditions of the target planting area, so as to improve the adaptability and feasibility of the mixed-species structure.

[0047] Furthermore, this invention employs heuristic optimization methods such as genetic algorithms or particle swarm optimization to determine the number of tree species to plant. This approach can quickly obtain near-optimal solutions under high-dimensional and nonlinear constraints, improving the stability and efficiency of the optimization process. By comparing the multi-objective optimization function values ​​after optimization, the invention automatically selects the tree species combination with the highest carbon sequestration efficiency, significantly enhancing the scientific validity, feasibility, and practical application value of the multi-tree species mixed carbon sequestration configuration scheme.

[0048] Example 2: As another preferred embodiment of the present invention, a differentiated mixed-species configuration is adopted for the selected tree species combination according to the terrain conditions of the target planting area, so as to improve the adaptability and feasibility of the mixed-species structure.

[0049] Specifically, when the target planting area is a relatively flat plain or a gently sloping area, a row-mixed planting method is preferred. For example... Figure 4 The diagram illustrates row-mixed planting, where different tree species are planted alternately along the same direction at predetermined row spacing, creating a regular horizontal distribution. Trees identified by the same symbol in the diagram represent the same species. This row-mixed planting method not only facilitates mechanized operations and subsequent management but also enables different tree species to complement each other's use of light, water, and nutrient resources at the row level, reducing the risks of pests and diseases and growth associated with continuous planting of a single species.

[0050] When the target planting area is a mountainous region with significant topographic relief, marked slope changes, or complex spatial conditions, a block-type mixed planting method is preferred. For example... Figure 5 The diagram illustrates a block-based mixed planting approach. Based on topographical undulations and plot boundaries, the target planting area is divided into several relatively independent smaller plots. Each smaller plot is planted with the same tree species, while different species are planted in different smaller plots. This block-based mixed planting method effectively utilizes the mountainous terrain, facilitating the selection of suitable tree species for planting, reducing the construction difficulty of mixed planting in complex terrain conditions, and improving the stability and feasibility of the overall afforestation implementation.

[0051] It should be noted that the sequence numbers of the above embodiments of the present invention are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or method. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0052] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0053] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence, characterized in that, The method includes: S1: Collect multi-source growth data of different tree species, and preprocess the multi-source growth data to obtain preprocessed multi-source growth data. Based on the preprocessed multi-source growth data, construct the individual growth characteristics and mixed growth characteristics of the tree species. S2: Based on the individual growth characteristics and mixed growth characteristics of the tree species, the carbon sequestration prediction model is used to predict the carbon sequestration values ​​of different tree species combinations under a given growth environment. S3: Based on the individual growth characteristics and mixed growth characteristics of the tree species, calculate the joint growth adaptability and functional complementarity of different tree species combinations. Combine the predicted carbon sequestration values ​​of the different tree species combinations under a given growth environment, and give a comprehensive score to the tree species combinations. Based on the comprehensive score, select a set of candidate tree species combinations. S4: Using the planting area of ​​tree species as a constraint, construct a multi-objective optimization function for the candidate tree species combination in the candidate tree species combination set. By solving the multi-objective optimization function, optimize and screen to obtain the candidate tree species combination with the highest carbon sequestration efficiency.

2. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 1, characterized in that, Step S1 involves collecting multi-source growth data from different tree species and preprocessing the multi-source growth data, including: The multi-source growth data includes growth characteristic data of tree species when they reach the estimated growth stage, historical carbon sequestration data, growth environment factor data, and operating cost data; The growth characteristic data include the average tree height, average diameter at breast height (DBH), and average root depth of the tree species. The historical carbon sequestration data includes the change in biomass of tree species per unit time and historical carbon sink monitoring records. The growth environment factor data includes the optimal environmental factor index values ​​for tree species growth of different types of environmental factors and the actual environmental factor index values ​​of the target planting area under different types of environmental factors, wherein the types of environmental factors include light, water and nutrients. The operating cost data includes the cost of purchasing seedlings per unit planting area, the cost of planting labor, and the cost of maintenance and management when the tree species reaches the estimated growth stage. The multi-source growth data is preprocessed, wherein the preprocessing procedure includes: S11: Complete missing data using interpolation or historical mean; S12: Normalize the growth data of different dimensions to obtain preprocessed multi-source growth data.

3. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 2, characterized in that, Based on the preprocessed multi-source growth data, individual growth characteristics and mixed growth characteristics of tree species are constructed, including: S13: Based on the preprocessed multi-source growth data, calculate the individual growth characteristics of the tree species, wherein the individual growth characteristics include growth potential characteristics, carbon sequestration efficiency characteristics, environmental adaptability characteristics, and cost constraint characteristics. S14: Based on the individual growth characteristics of the tree species, introduce the interaction relationship between tree species to construct the mixed growth characteristics between any two tree species. The mixed growth characteristics include growth complementarity characteristics, competition inhibition characteristics, and stability synergy characteristics. Growth complementarity characteristics describe the degree of complementarity between different tree species in spatial structure or resource utilization. Competition inhibition characteristics describe the competition relationship between tree species for light, water, and nutrients. Stability synergy characteristics describe the impact of the mixed structure on community stability and resistance to disturbance.

4. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 1, characterized in that, The carbon fixation prediction model in step S2 includes a growth environment coding module, a single-unit feature coding module, a mixed growth feature coding module, and a feature fusion and nonlinear mapping module, including: The growth environment coding module is used to perform nonlinear mapping on the actual environmental factor index values ​​of the target planting area to form an environmental embedding representation that can characterize the overall growth conditions of the area. The individual feature encoding module is used to encode the growth potential, carbon sequestration efficiency, environmental adaptability and cost constraint characteristics of each tree species in the tree species combination, and to use an attention mechanism to weight and fuse the encoding results of different tree species to highlight the individual features of tree species that contribute more to the current growth environment. The mixed growth feature encoding module is used to encode and aggregate the growth complementarity features, competition inhibition features, and stability synergy features between any two tree species in the tree species combination, so as to characterize the interaction relationship under mixed tree species conditions; The feature fusion and nonlinear mapping module is used to splice and fuse environmental embedding representations, single-unit encoded features and hybrid encoded features, and to achieve adaptive adjustment of different feature information through a multi-layer nonlinear mapping structure based on a gating mechanism, and output the predicted carbon sequestration value of the tree species combination under a given growth environment.

5. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 4, characterized in that, Step S2, which uses a carbon sequestration prediction model to predict the carbon sequestration values ​​of different tree species combinations under a given growth environment, also includes: S21: Divide the tree species into multiple tree species combinations according to the preset range of mixed tree species; S22: Based on the individual growth characteristics and mixed growth characteristics of the tree species, extract the individual growth characteristics of each tree species in the tree species combination and the mixed growth characteristics between any two tree species. S23: The actual environmental factor index values ​​of the target planting area under different types of environmental factors are taken as the given growth environment. The growth environment coding module performs nonlinear mapping on the given growth environment to obtain the environmental embedding representation. S24: The individual feature encoding module encodes the individual growth features of each tree species in the tree species combination, and fuses the encoding results based on the attention mechanism to obtain the individual encoded features of the tree species combination. S25: The mixed growth feature encoding module encodes and fuses the mixed growth features in the tree species combination to obtain the mixed encoding features of the tree species combination; S26: The feature fusion and nonlinear mapping module concatenates the environment embedding representation, the individual coding features of the tree species combination, and the hybrid coding features of the tree species combination, and outputs the predicted carbon sequestration value of the tree species combination using a multi-layer nonlinear mapping method based on a gating mechanism.

6. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 1, characterized in that, The S3 step calculates the combined growth adaptability and functional complementarity of different tree species combinations, including: S31: Calculate the contribution weight coefficient of the tree species in the combination of tree species based on the individual growth characteristics and mixed growth characteristics of the tree species. S32: Based on the contribution weight coefficient of the tree species, perform weighted mapping on the individual growth characteristics of the tree species in the tree species combination to obtain the joint growth adaptability of the tree species combination; S33: Based on the contribution weight coefficient of the tree species, calculate the comprehensive contribution weight coefficient of any two tree species in the tree species combination, perform weighted mapping on the mixed growth characteristics between tree species, and obtain the functional complementarity of the tree species combination.

7. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 6, characterized in that, Step S3, which combines the predicted carbon sequestration values ​​of different tree species combinations under a given growth environment to perform a comprehensive evaluation of the tree species combinations, also includes: S34: Obtain the predicted carbon sequestration value of the tree species combination under a given growth environment; S35: Based on the combined growth adaptability, functional complementarity, and predicted carbon sequestration of the tree species combination, a comprehensive score is given to the tree species combination using a joint scoring function; S36: Based on the comprehensive score, select the top 10% of tree species combinations with the highest comprehensive scores as candidate tree species combinations, and construct a candidate tree species combination set.

8. The method for optimizing and screening carbon sequestration through multi-species mixed intercropping based on artificial intelligence as described in claim 1, characterized in that, In step S4, a multi-objective optimization function is constructed for the candidate tree species combinations in the candidate tree species combination set, using the planting area of ​​the tree species as a constraint. The expression of the multi-objective optimization function is as follows: ； ； in, Represents a multi-objective optimization function. Indicates candidate tree species combinations The multi-objective optimization function value, Indicates candidate tree species combinations The predicted value of carbon sequestration, Indicates candidate tree species combinations Overall score This represents the maximum value among the comprehensive scores of all candidate tree species combinations in the candidate tree species combination set. Indicates candidate tree species combinations The d-th tree species The number of plants, Indicates candidate tree species combinations The d-th tree species The planting area of ​​a single tree This indicates the arable area of ​​the target planting region. This indicates the minimum number of tree species to be planted. Indicates candidate tree species combinations The number of tree species, including candidate tree species combinations The number of trees planted for each species is the parameter to be solved in the multi-objective optimization function; This indicates a constraint.

Images