Forest seedling raising period management optimization method based on adaptive algorithm

Abstract

The invention discloses a forest seedling raising period management optimization method based on an adaptive algorithm. The method comprises the following steps: S1, collecting multi-source environment data and seedling growth state data in a forest seedling raising process; s2, the original index data set is processed, a standardized index matrix is obtained, and a stage index evaluation input set is constructed; s3, constructing a multi-target periodic scheduling model based on the stage index evaluation input set; s4, introducing an entropy weight adaptive grey wolf optimization algorithm to solve the multi-target scheduling function set, and outputting a periodic scheduling optimal solution; s5, generating a seedling culture management execution strategy based on the period scheduling optimal solution; s6, forming a feedback monitoring data set by the environment feedback data collected in real time and the seedling feedback data, and generating an updated seedling management execution strategy; and S7, synchronizing the updated seedling culture management execution strategy to the seedling culture management platform. According to the invention, intelligent scheduling and adaptive management optimization of the whole forest seedling raising process are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agricultural management, and in particular to a forest seedling cycle management optimization method based on an adaptive algorithm. Background Art

[0002] In the context of sustainable forest resource management, forest seedling cultivation, as the front-end link of afforestation operations, has a cycle management quality that directly affects the subsequent forest growth efficiency and ecological value realization. At present, the cycle management of forest seedling cultivation mostly adopts a method that relies on manual experience to formulate plans and adjusts them in combination with periodic inspection data. Traditional seedling management strategies are usually based on the planting season, regional climate and empirical rules to formulate operation time and resource allocation standards for several key nodes, and manually observe and record changes in seedling growth status and environmental conditions during the seedling cultivation process to assist in determining whether to adjust the management plan. Although this method has certain practicality in a relatively stable natural environment, its scheduling accuracy and resource matching efficiency are significantly insufficient when dealing with complex climate changes, increased soil heterogeneity or changes in seedling density, which often leads to delays in the seedling cultivation cycle, waste of resources and fluctuations in seedling quality.

[0003] In recent years, with the widespread adoption of environmental monitoring equipment, the agricultural Internet of Things (IoT), and data collection terminals, data-driven management approaches have gradually been introduced into seedling management. Some systems have attempted to utilize environmental parameter acquisition equipment and image recognition technology to obtain basic indicators such as seedling height, soil moisture, and light intensity, which are then used to construct growth models or early warning systems for the seedling stage. Building on this foundation, some studies have proposed using optimization algorithms to optimize water and fertilizer management or labor inputs during the seedling process. Common methods include genetic algorithms, ant colony algorithms, and fuzzy control strategies. While these approaches have achieved some success in single-objective optimization or stage-by-stage scheduling, they still face limitations when addressing the multi-objective, multi-stage scheduling problems involved in the entire seedling process.

[0004] Existing multi-objective scheduling optimization methods face three main problems in forest seedling cultivation scenarios: First, most optimization methods fail to simultaneously take into account the three core indicators of minimizing seedling cultivation time, minimizing resource utilization, and optimizing seedling quality. There is a lack of effective coordination mechanism between optimization objectives, resulting in scheduling results biased towards a single objective, making it difficult to balance the comprehensive benefits in actual production; Second, current optimization methods usually use fixed weight functions or empirical weighting methods to construct objective functions, and lack a mechanism for dynamically adjusting weights based on data fluctuations, resulting in insufficient sensitivity of the optimization model to different seedling cultivation stages and difficulty in adapting to the changing characteristics of indicator weights as the stages evolve; Third, existing optimization scheduling schemes are mostly static one-time generation strategies, which cannot be dynamically optimized and adjusted based on real-time feedback during implementation. When the environment or growth status deviates, the system lacks a response mechanism, making it difficult to timely correct the cycle schedule and resource allocation strategy.

[0005] Therefore, how to provide a forest seedling cycle management optimization method based on an adaptive algorithm is an urgent problem that those skilled in the art need to solve. Summary of the Invention

[0006] One object of the present invention is to propose a method for optimizing forest seedling cycle management based on an adaptive algorithm. The present invention makes full use of multi-source environmental data and seedling growth status data, and through standardized processing and phased indicator extraction, constructs a multi-objective cycle scheduling model for the entire seedling process. The model takes the shortest total seedling time, the minimum resource consumption and the maximum seedling health score as optimization objectives, integrates the entropy weight calculation mechanism and the gray wolf optimization algorithm, dynamically generates a multi-objective fusion fitness function, and obtains the optimal solution for cycle scheduling by guiding individual iterative search. At the same time, the present invention designs a scheduling strategy structure mapping mechanism that can convert the optimal solution into a phased task schedule, resource allocation plan and instruction sequence, and has deployability and system compatibility. During the actual execution process, deviation monitoring is carried out by collecting feedback data in real time, and the scheduling model reconstruction is automatically triggered to achieve dynamic optimization response. This method has the advantages of strong target coordination, high scheduling accuracy, reasonable resource allocation and strong environmental adaptability, and significantly improves the intelligence level and production efficiency of forest seedling management.

[0007] A forest seedling cycle management optimization method based on an adaptive algorithm according to an embodiment of the present invention includes the following steps:

[0008] S1. Collect multi-source environmental data and seedling growth status data during forest seedling cultivation to generate a set of original indicator data;

[0009] S2. Process the original indicator data set to obtain a standardized indicator matrix, and construct a stage indicator evaluation input set corresponding to the standardized indicator matrix in combination with the seedling stage labels;

[0010] S3. Construct a multi-objective periodic scheduling model based on the phased indicator evaluation input set and output a multi-objective scheduling function set;

[0011] S4. Introduce the entropy-weighted adaptive grey wolf optimization algorithm to solve the multi-objective scheduling function set. By calculating the information entropy of each indicator, the entropy weight vector is generated, the fusion objective function is constructed and iteratively updated, and the optimal solution for periodic scheduling is output.

[0012] S5. Generate a seedling management execution strategy based on the optimal solution of periodic scheduling;

[0013] S6. The environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, the feedback monitoring data set is compared with the original indicator data set, an updated periodic scheduling optimal solution is output, and an updated seedling management execution strategy is generated;

[0014] S7. Synchronize the seedling management execution strategy or the updated seedling management execution strategy to the seedling management platform to complete the implementation deployment and task scheduling execution of the cycle optimization results.

[0015] Optionally, the original indicator data set includes environmental parameters and seedling growth status parameters; the environmental parameters include temperature, humidity, light intensity, soil moisture content and soil nutrient concentration; the seedling growth status parameters include seedling height, growth rate, stem diameter, leaf area index, root development index, root-crown ratio, physiological maturity score and disease and insect pest incidence.

[0016] Optionally, the S2 specifically includes:

[0017] S21. Perform data cleaning operations on the data of each dimension in the original indicator data set, remove missing values, duplicate values, and abnormal data exceeding the statistical threshold, and generate a cleaned indicator data set;

[0018] S22. Perform normalization processing on each dimension indicator in the cleaned indicator data set, calculate the normalized result of each indicator value using a linear normalization method, arrange all normalized results by sample, and construct a normalized indicator matrix;

[0019] S23. According to the normalized values ​​of seedling height, growth rate, and root development index corresponding to each sample in the standardized indicator matrix, a threshold judgment rule for seedling stage identification is set, and the stage to which each sample belongs is automatically determined based on the threshold judgment rule to generate a seedling stage label set;

[0020] S24. Combine the standardized indicator matrix with the seedling stage label set to construct a stage indicator evaluation input set.

[0021] Optionally, the S3 specifically includes:

[0022] S31, constructing a multi-objective periodic scheduling model, wherein the multi-objective periodic scheduling model includes an indicator input fusion layer, a stage duration evaluation layer, a resource usage estimation layer, and a health score aggregation layer;

[0023] S32. In the indicator input fusion layer, each standardized indicator vector in the stage indicator evaluation input set is input into the fusion layer, and dimension uniform mapping, feature reconstruction and normalization processing are performed to generate a fusion input feature set;

[0024] S33. In the stage duration evaluation layer, based on the fused input feature set and the corresponding seedling stage label, the duration estimate of each seedling stage is calculated respectively, and the duration estimates of all seedling stages are accumulated to obtain the total seedling duration objective function;

[0025] S34. In the resource usage estimation layer, based on the resource-related indicators in the fused input features, the resource usage of each type involved in each seedling stage is counted separately, and the resource usage of all resource types in all seedling stages is summed up to construct a resource consumption objective function;

[0026] S35. In the health score aggregation layer, extract the three dimensions of seedling height, growth rate, and root development index from the fused input features, calculate the health score value of each sample separately, and sum the health scores of all samples in all seedling stages and divide them by the total number of samples to obtain the seedling health score objective function;

[0027] S36. The total seedling raising time objective function, the resource consumption objective function and the seedling health score objective function are combined to form a multi-objective scheduling function set F = {f1, f2, f3}.

[0028] Optionally, the S4 specifically includes:

[0029] S41, initializing the search space dimension of the gray wolf optimization algorithm, determining the search dimension according to the input parameter structure of the multi-objective periodic scheduling function set, and randomly generating initial position vectors of multiple gray wolf individuals to form an initial population;

[0030] S42. Calculate the fitness value of each individual gray wolf under the objective function of total seedling cultivation time, the objective function of resource consumption, and the objective function of seedling health score, and calculate the information entropy for each objective function based on the normalized fitness of all individuals under each objective function;

[0031] S43. Calculate the entropy weight according to the information entropy corresponding to each objective function. The smaller the information entropy, the greater the weight of the objective function. The sum of the entropy weights of all objective functions is 1.

[0032] S44, weighted multiplication of the entropy weight of each objective function and the corresponding individual fitness value, and summing the results over the three objective function dimensions to obtain the fusion fitness value corresponding to each gray wolf individual;

[0033] S45. In each iteration, the gray wolf individuals are sorted according to the fusion fitness value, and the currently optimal, second-optimal, and third-optimal individuals are selected as guide individuals. The new positions of all gray wolf individuals are calculated according to the position update mechanism of the standard gray wolf optimization algorithm.

[0034] S46. Repeat the fitness calculation, entropy weight update and position update processes until the preset maximum number of iterations is reached, and finally output the individual position vector with the optimal fusion fitness value. The parameter solution corresponding to the optimal individual position vector is the optimal solution for periodic scheduling.

[0035] Optionally, calculating the information entropy includes normalizing the fitness value of each gray wolf individual under multiple objective functions, constructing a normalized fitness sequence for each objective function, and calculating the information entropy of the j-th objective function according to the information entropy formula:

[0036]

[0037] Among them, e j represents the information entropy value of the jth objective function in the multi-objective scheduling function set, p ij represents the normalized fitness value of the i-th gray wolf individual under the j-th objective function, n represents the number of individuals in the gray wolf population, k represents the normalization constant, ε represents a positive constant to prevent the logarithmic term from being zero, i represents the index number of the gray wolf population individual, and j represents the index number of the objective function in the multi-objective scheduling function set.

[0038] Optionally, the S45 specifically includes:

[0039] S451, the position update mechanism of the standard gray wolf optimization algorithm includes guiding individual selection, prey position estimation, guiding vector construction and position fusion update;

[0040] S452, guide individual selection: all gray wolf individuals are ranked based on the fusion fitness value, and the top three individuals with the best fitness values ​​are selected, which are defined as guide individuals α, β and δ respectively;

[0041] S453. Prey Position Estimation: For any individual i in the gray wolf population, calculate the directional distance vector of any individual i relative to the guide individual α, β, and δ, and generate the perturbation coefficient and position offset based on the current search round to estimate the target prey position of any individual i in the three directions;

[0042] S454: Execute a guidance position vector construction operation, and construct a fitted position update expression for each direction guidance result based on the three estimated positions and the corresponding distance control vectors;

[0043] S455. Perform a position fusion update operation, perform a mean fusion operation on the position vectors obtained by guiding in the three directions, and obtain the new position vector of the i-th gray wolf individual in this round of iteration as the update result of the current search state.

[0044] Optionally, the S6 specifically includes:

[0045] S51. Express the optimal solution of periodic scheduling as a structure vector where s k represents the scheduling time and resource allocation results of the kth seedling raising stage, where k is the total number of seedling raising stages;

[0046] S52, according to the structure vector The scheduling time of each stage in the construction phase task execution schedule is a two-dimensional array T = {t kj}, where t kj represents the planned execution time of the jth task in the kth stage;

[0047] S53, according to the structure vector The resource consumption components of each stage in the process are used to construct a resource allocation configuration table, which is a three-dimensional structure R={r kjm}, where r kjm It represents the amount of resources of type m allocated to type j management action in the kth stage. The resource types include water, fertilizer, light and manpower.

[0048] S54, generating a management action instruction sequence for each stage based on the stage task execution schedule T and the resource allocation configuration table R, wherein the instruction sequence is a management action instruction sequence set;

[0049] S55. Integrate the task execution schedule T, the resource allocation configuration table R and the management action instruction sequence set to form a complete seedling management execution strategy.

[0050] Optionally, the S6 specifically includes:

[0051] S61. During the implementation of the seedling management execution strategy, the environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, and feature dimension alignment and time alignment are performed with the original indicator data set;

[0052] S62. Perform deviation calculation on the key indicators of the same dimension in the feedback monitoring data set and the original indicator data set to obtain an indicator deviation vector Δ(t);

[0053] S63, compare all the index items in the index deviation vector Δ(t) one by one with the preset threshold vector θ=[θ1,θ2,…,θ i ,…,θ m ] to compare, if there is any one that satisfies Δ (t) ≥θ i , it is determined that the current feedback monitoring data has triggered an abnormal deviation;

[0054] S64. When the deviation judgment result is a trigger state, the multi-objective periodic scheduling model is reconstructed using the phased indicator evaluation input set and the entropy weight adaptive grey wolf optimization algorithm, the updated periodic scheduling optimal solution is output, and the updated seedling management execution strategy is generated.

[0055] Optionally, the S7 specifically includes:

[0056] S71, receiving a seedling management execution strategy or an updated seedling management execution strategy, converting the stage task execution schedule T into a set of structured task scheduling entries, each entry encapsulating a task identifier, a target stage, an execution time, and a required resource category label in the form of a key-value pair, to form a time scheduling mapping set;

[0057] S72, mapping the resource allocation configuration table R to a set of allocation instructions, each allocation instruction is a structure triple<k,j,m> →r kjm Indicates the allocation quantity for the corresponding phase, task, and resource type;

[0058] S73, serialize and encode the management action instruction sequence set into an action control instruction stream, and generate a control instruction format set that matches the communication interface of each physical control node;

[0059] S74: The scheduling mapping set, the allocation instruction set, and the control instruction format set are aggregated and encapsulated into a policy synchronization object, and a policy deployment request is initiated to the seedling management platform to complete the policy push of the periodic scheduling result;

[0060] S75. The seedling management platform parses the strategy synchronization object and sends each task scheduling entry, resource allocation item and control instruction therein to each scheduling control sub-module in the platform to complete the system-level deployment of the seedling management execution strategy and task scheduling execution initialization.

[0061] The beneficial effects of the present invention are:

[0062] This paper collects multi-source environmental data and seedling growth status data from the forest seedling cultivation process, establishes a structured set of original indicator data, and builds a standardized indicator matrix and a set of phased indicator evaluation inputs on this basis, thus achieving quantitative modeling of key influencing factors at each stage of seedling cultivation. Furthermore, by constructing a multi-objective periodic scheduling model, three objective functions are set: total seedling cultivation time, resource consumption, and seedling health score. The effectiveness of seedling cultivation management is measured from the three dimensions of time efficiency, resource utilization, and growth quality. This helps to systematically reveal the inherent relationship between task execution, resource allocation, and seedling growth at each stage, thereby enhancing the goal-orientedness and scientific rationality of the management strategy.

[0063] In order to solve the problem of difficulty in coordinating weights in the multi-objective optimization process, the present invention introduces an entropy weight mechanism based on information entropy calculation and integrates the gray wolf optimization algorithm to solve the scheduling function. This combination method can dynamically adjust the target weights according to the fitness dispersion of each objective function in the gray wolf population, making the scheduling process more adaptable and with stronger target coordination capabilities. The gray wolf optimization algorithm uses fused fitness values ​​to guide individual search paths during the iteration process, and combines multi-directional guidance with position fusion strategies to improve the global convergence ability and local search accuracy of the scheduling parameter solution. The final output of the optimal solution for periodic scheduling not only meets the multi-objective balance requirements, but also has high execution stability and model adaptability.

[0064] On this basis, the present invention also establishes a structured mapping mechanism from optimization results to seedling management execution strategies, converts the optimal scheduling solution into a stage task schedule, a resource allocation configuration table and a management action instruction sequence, and constructs a deviation judgment and rescheduling trigger mechanism through real-time collection of feedback data, thereby realizing closed-loop management of the entire process. When key indicators of environmental parameters or growth status deviate during system operation, the model can automatically start the scheduling reconstruction process and output the updated execution strategy, thereby maintaining the stability and flexibility of seedling cycle control. Overall, the present invention realizes a system solution that integrates data modeling, optimization solution, strategy deployment and dynamic response in forest seedling cycle management, effectively improving the scheduling efficiency of seedling operations, the rationality of resource utilization and the quality assurance of seedlings. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0066] Figure 1 This is a flow chart of a forest seedling cycle management optimization method based on an adaptive algorithm proposed in the present invention;

[0067] Figure 2 This is a structural diagram of a multi-objective periodic scheduling model in a forest seedling periodic management optimization method based on an adaptive algorithm proposed in the present invention;

[0068] Figure 3 This is a schematic diagram of the iterative optimization process of the entropy weight adaptive gray wolf optimization algorithm in the forest seedling cycle management optimization method based on the adaptive algorithm proposed in the present invention. DETAILED DESCRIPTION

[0069] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0070] refer to Figure 1-3 , a forest seedling cycle management optimization method based on an adaptive algorithm, comprising the following steps:

[0071] S1. Collect multi-source environmental data and seedling growth status data during forest seedling cultivation to generate a set of original indicator data;

[0072] S2. Process the original indicator data set to obtain a standardized indicator matrix, and construct a stage indicator evaluation input set corresponding to the standardized indicator matrix in combination with the seedling stage labels;

[0073] S3. Construct a multi-objective periodic scheduling model based on the phased indicator evaluation input set and output a multi-objective scheduling function set;

[0074] S4. Introduce the entropy-weighted adaptive grey wolf optimization algorithm to solve the multi-objective scheduling function set. By calculating the information entropy of each indicator, the entropy weight vector is generated, the fusion objective function is constructed and iteratively updated, and the optimal solution for periodic scheduling is output.

[0075] S5. Generate a seedling management execution strategy based on the optimal solution of periodic scheduling;

[0076] S6. The environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, the feedback monitoring data set is compared with the original indicator data set, an updated periodic scheduling optimal solution is output, and an updated seedling management execution strategy is generated;

[0077] S7. Synchronize the updated seedling management execution strategy to the seedling management platform to complete the implementation deployment and task scheduling execution of the cycle optimization results.

[0078] The present invention proposes a method for optimizing the cycle management of forest seedlings based on an adaptive algorithm. By constructing a standardized indicator system based on multi-source environmental data and seedling growth status data, and combining a multi-objective cycle scheduling model with an entropy-weighted adaptive gray wolf optimization algorithm, the scheduling optimization and refined management of resource allocation for the entire process of forest seedling cultivation are achieved. By introducing a dynamic weighting mechanism of information entropy, the coordination and adaptability of multi-objective optimization are improved, and the optimization results can be mapped into task schedules, resource allocation tables, and instruction sequences, which are executable and deployable. The system supports real-time feedback-triggered rescheduling, has dynamic adjustment capabilities, and can effectively respond to environmental fluctuations and abnormal growth problems. The overall execution efficiency of seedling management, resource utilization efficiency, and seedling health level are improved.

[0079] In this embodiment, the original indicator data set includes environmental parameters and seedling growth status parameters; the environmental parameters include temperature, humidity, light intensity, soil moisture content and soil nutrient concentration; the seedling growth status parameters include seedling height, growth rate, stem diameter, leaf area index, root development index, root-crown ratio, physiological maturity score and disease and insect pest incidence.

[0080] This method collects temperature, humidity, light intensity, soil moisture content, and soil nutrient concentration as environmental parameters, as well as seedling height, growth rate, stem diameter, leaf area index, root development index, root-to-shoot ratio, physiological maturity score, and pest and disease incidence as growth status parameters. This constructs a comprehensive set of raw indicator data, providing a high-dimensional input foundation for subsequent scheduling models. This data system helps improve the accuracy of environmental adaptability analysis and seedling growth status assessment during seedling management, enhancing the scientific nature and accuracy of scheduling optimization.

[0081] In this embodiment, S2 specifically includes:

[0082] S21. Perform data cleaning operations on the data of each dimension in the original indicator data set, remove missing values, duplicate values, and abnormal data exceeding the statistical threshold, and generate a cleaned indicator data set;

[0083] S22. Perform normalization processing on each dimension indicator in the cleaned indicator data set, calculate the normalized result of each indicator value using a linear normalization method, arrange all normalized results by sample, and construct a normalized indicator matrix;

[0084] S23. According to the normalized values ​​of seedling height, growth rate, and root development index corresponding to each sample in the standardized indicator matrix, a threshold judgment rule for seedling stage identification is set, and the stage to which each sample belongs is automatically determined based on the threshold judgment rule to generate a seedling stage label set;

[0085] S24. Combine the standardized indicator matrix with the seedling stage label set to construct a stage indicator evaluation input set.

[0086] This method improves data quality and model input stability by cleaning and linearly normalizing raw indicator data, removing missing, duplicate, and outlier values. Threshold rules are set based on seedling height, growth rate, and root development index to automatically identify and label each seedling stage. Furthermore, a set of stage-specific indicator evaluation inputs is constructed to provide accurate stage-specific support for the scheduling model, enhancing the targeted nature and stage-specific adaptability of the scheduling process.

[0087] In this embodiment, S3 specifically includes:

[0088] S31, constructing a multi-objective periodic scheduling model, wherein the multi-objective periodic scheduling model includes an indicator input fusion layer, a stage duration evaluation layer, a resource usage estimation layer, and a health score aggregation layer;

[0089] S32. In the indicator input fusion layer, each standardized indicator vector in the stage indicator evaluation input set is input into the fusion layer, and dimension uniform mapping, feature reconstruction and normalization processing are performed to generate a fusion input feature set;

[0090] S33. In the stage duration evaluation layer, based on the fused input feature set and the corresponding seedling stage label, the duration estimate of each seedling stage is calculated respectively, and the duration estimate values ​​of all seedling stages are accumulated to obtain the total seedling duration objective function, where the total seedling duration is equal to the sum of the duration values ​​of all seedling stages;

[0091] S34. In the resource usage estimation layer, based on the resource-related indicators in the fused input features, the resource usage of each type involved in each seedling stage is counted separately, and the resource usage of all resource types in all seedling stages is summed up to construct a resource consumption objective function. The resource consumption objective function represents the cumulative total consumption of all resources in all seedling stages;

[0092] S35. In the health score aggregation layer, extract the three dimensions of seedling height, growth rate, and root development index from the fused input features, calculate the health score value for each sample separately, and sum the health scores of all samples in all seedling stages and divide them by the total number of samples to obtain the seedling health score objective function. The seedling health score objective function represents the average health score of all samples in the seedling cycle;

[0093] S36. The total seedling raising time objective function, the resource consumption objective function and the seedling health score objective function are combined to form a multi-objective scheduling function set F = {f1, f2, f3}.

[0094] The present invention realizes the unified modeling of time control, resource allocation and seedling quality in the seedling cycle by constructing a multi-objective periodic scheduling model including an indicator input fusion layer, a stage duration evaluation layer, a resource usage estimation layer and a health score aggregation layer. In the input fusion layer, feature reconstruction and normalization operations are used to improve the indicator dimension consistency and feature expression ability, providing a stable input for subsequent modeling. In the duration evaluation layer and the resource estimation layer, the duration and resource consumption of each seedling stage are calculated respectively, and the total seedling duration and resource consumption objective function are constructed; in the health score aggregation layer, seedling height, growth rate and root development index are used as core indicators to quantify the seedling quality performance. The final multi-objective scheduling function set can provide a clear evaluation basis for the optimization algorithm, and significantly improve the comprehensive balance ability of the scheduling strategy in terms of efficiency, resource utilization and seedling health.

[0095] In this embodiment, the S4 specifically includes:

[0096] S41, initializing the search space dimension of the gray wolf optimization algorithm, determining the search dimension according to the input parameter structure of the multi-objective periodic scheduling function set, and randomly generating initial position vectors of multiple gray wolf individuals to form an initial population;

[0097] S42. Calculate the fitness value of each individual gray wolf under the objective function of total seedling cultivation time, the objective function of resource consumption, and the objective function of seedling health score, and calculate the information entropy for each objective function based on the normalized fitness of all individuals under each objective function;

[0098] S43. Calculate the entropy weight according to the information entropy corresponding to each objective function. The smaller the information entropy, the greater the weight of the objective function. The sum of the entropy weights of all objective functions is 1.

[0099] S44, weighted multiplication of the entropy weight of each objective function and the corresponding individual fitness value, and summing the values ​​across the three objective function dimensions to obtain the fusion fitness value corresponding to each gray wolf individual, which is used to measure the comprehensive performance of the individual under multiple objectives;

[0100] S45. In each iteration, the gray wolf individuals are sorted according to the fusion fitness value, and the currently optimal, second-optimal, and third-optimal individuals are selected as guide individuals. According to the position update mechanism of the standard gray wolf optimization algorithm, the new positions of all gray wolf individuals are calculated. The new positions are the average estimated positions guided by the three optimal individuals, which are used to balance local development and global search.

[0101] S46. Repeat the fitness calculation, entropy weight update and position update processes until the preset maximum number of iterations is reached, and finally output the individual position vector with the optimal fusion fitness value. The parameter solution corresponding to the optimal individual position vector is the optimal solution for periodic scheduling, and the optimal solution for periodic scheduling is used to generate a seedling management execution strategy.

[0102] The present invention introduces an entropy-weighted adaptive gray wolf optimization algorithm to efficiently solve a set of multi-objective periodic scheduling functions. During the optimization process, a search space is first constructed based on the scheduling model parameters, and multiple gray wolf individuals are initialized to form the basis of the population. By calculating the fitness of each individual under three types of objective functions, namely, seedling raising time, resource consumption, and health score, and calculating the information entropy based on the normalized results, the weights of each objective function are dynamically generated to achieve an adaptive balance between the objectives. Subsequently, the entropy weight is weighted and fused with the individual fitness to form a comprehensive fitness value to guide the population iteration. Each round of iteration uses the best three individuals to jointly guide and perform position updates, which enhances the convergence efficiency of the search process and the stability of the solution. The optimal individual position vector finally output is used as the periodic scheduling parameter solution, providing a basis for the generation of management strategies. This method improves the coordination ability of multi-objective optimization and enhances the accuracy of scheduling solutions and global search capabilities.

[0103] In this embodiment, the information entropy calculation includes normalizing the fitness value of each gray wolf individual under multiple objective functions, constructing a normalized fitness sequence for each objective function, and calculating the information entropy of the j-th objective function according to the information entropy formula:

[0104]

[0105] Among them, e j represents the information entropy value of the jth objective function in the multi-objective scheduling function set, p ij represents the normalized fitness value of the i-th gray wolf individual under the j-th objective function, n represents the number of individuals in the gray wolf population, k represents the normalization constant, ε represents a positive constant to prevent the logarithmic term from being zero, i represents the index number of the gray wolf population individual, and j represents the index number of the objective function in the multi-objective scheduling function set.

[0106] This invention introduces an information entropy calculation mechanism into the multi-objective optimization process to dynamically assess the importance of each objective function. Specifically, the fitness values ​​of each individual in the gray wolf population under each objective function are normalized to construct a normalized fitness sequence corresponding to each objective function. Subsequently, the distribution of these normalized values ​​is analyzed to calculate the information entropy value of each objective function. Information entropy measures the degree of dispersion of the fitness distribution of a particular objective function within the entire population. The more concentrated the distribution of the normalized fitness values ​​of a particular objective function, the lower its information entropy value, indicating that the objective function has a stronger ability to discriminate the optimization results. When assigning weights, objective functions with lower information entropy values ​​are assigned greater weights. In this way, the system can adaptively adjust the influence of each objective function during the optimization process, avoiding reliance on fixed, manually set weights. This effectively improves the representativeness of the fused fitness value and the reliability of the optimized solution. This mechanism makes multi-objective optimization more flexible and data-driven, enhancing the scheduling model's adaptability to diverse optimization requirements.

[0107] In this embodiment, the S45 specifically includes:

[0108] S451, the position update mechanism of the standard gray wolf optimization algorithm includes guiding individual selection, prey position estimation, guiding vector construction and position fusion update;

[0109] S452, guide individual selection: all gray wolf individuals are ranked based on the fusion fitness value, and the top three individuals with the best fitness values ​​are selected, which are defined as guide individuals α, β and δ respectively;

[0110] S453. Prey Position Estimation: For any individual i in the gray wolf population, calculate the directional distance vector of any individual i relative to the guide individual α, β, and δ, and generate the perturbation coefficient and position offset based on the current search round to estimate the target prey position of any individual i in the three directions;

[0111] S454: Execute a guidance position vector construction operation, and construct a fitted position update expression for each direction guidance result based on the three estimated positions and the corresponding distance control vectors;

[0112] S455. Perform a position fusion update operation, perform a mean fusion operation on the position vectors obtained by guiding in the three directions, and obtain the new position vector of the i-th gray wolf individual in this round of iteration as the update result of the current search state.

[0113] The present invention improves the synergistic efficiency of the global search capability and local development capability of the optimization process by refining the position update mechanism of the gray wolf optimization algorithm. In each round of iteration, the individuals in the population are sorted according to the fusion fitness value, and the top three individuals in fitness are selected as guide individuals, and the prey positions of any individual relative to these three are estimated separately. After the position is estimated by combining the disturbance coefficient and the direction offset, the new individual position is obtained by constructing a fitting vector and performing a mean fusion operation. This mechanism implements a position update strategy under multiple guiding directions, enhances the adaptability to complex scheduling search space, and improves the stability of the optimization results and the probability of obtaining the global optimal solution.

[0114] In this embodiment, the S5 specifically includes:

[0115] S51. Express the optimal solution of periodic scheduling as a structure vector where s k represents the scheduling time and resource allocation results of the kth seedling raising stage, where k is the total number of seedling raising stages;

[0116] S52, according to the structure vector The scheduling time of each stage in the construction phase task execution schedule is a two-dimensional array T = {t kj}, where t kj represents the planned execution time of the jth task in the kth stage;

[0117] S53, according to the structure vector The resource consumption components of each stage in the process are used to construct a resource allocation configuration table, which is a three-dimensional structure R={r kjm}, where r kjm It represents the amount of resources of type m allocated to type j management action in the kth stage. The resource types include water, fertilizer, light and manpower.

[0118] S54, generating a management action instruction sequence for each stage based on the stage task execution schedule T and the resource allocation configuration table R, wherein the instruction sequence is a management action instruction sequence set;

[0119] S55. Integrate the task execution schedule T, resource allocation configuration table R and management action instruction sequence set to form a complete seedling management execution strategy, which is used to guide the planning implementation and management execution of seedling operations in subsequent stages.

[0120] The present invention represents the optimal solution for periodic scheduling as a structural vector, extracts the scheduling time and resource allocation results of each seedling raising stage, constructs a task execution schedule and a resource allocation configuration table, and respectively describes the execution time and resource usage of each task in each stage. The task schedule is a two-dimensional array that clarifies the specific execution time of each management task; the resource allocation table accurately records the usage of various resources in each stage, including water, fertilizer, light and manpower. The schedule and resource allocation data are further mapped into a sequence of management action instructions to form a complete set of management execution strategies for guiding the subsequent seedling raising operation plans. This strategy has a high degree of structure and can achieve unified control of the time, resources and actions of management tasks, thereby improving the operability and execution efficiency of seedling raising scheduling.

[0121] In this embodiment, S6 specifically includes:

[0122] S61. During the implementation of the seedling management execution strategy, the environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, and feature dimension alignment and time alignment are performed with the original indicator data set;

[0123] S62. Perform deviation calculation on the key indicators of the same dimension in the feedback monitoring data set and the original indicator data set to obtain an indicator deviation vector Δ(t);

[0124] S63, compare all the index items in the index deviation vector Δ(t) one by one with the preset threshold vector θ=[θ1,θ2,…,θ i ,…,θ m ] to compare, if there is any one that satisfies Δ (t) ≥θ i , it is determined that the current feedback monitoring data has triggered an abnormal deviation;

[0125] S64. When the deviation judgment result is a trigger state, the multi-objective periodic scheduling model is reconstructed using the phased indicator evaluation input set and the entropy weight adaptive grey wolf optimization algorithm, and the updated periodic scheduling optimal solution is output. Based on the updated periodic scheduling optimal solution, an updated seedling management execution strategy is generated.

[0126] The present invention introduces a real-time feedback mechanism in the process of seedling management execution, forms a feedback monitoring data set by collecting environmental and seedling status data, and performs feature dimension and time alignment processing with the original indicator data. The system performs deviation calculation on key indicators, generates indicator deviation vectors, and compares them item by item with the set threshold vectors. If any indicator exceeds the threshold, it is judged as an abnormal state, triggering the reconstruction of the scheduling model. Subsequently, the multi-objective periodic scheduling model is updated using the current stage indicator evaluation data, and a new optimal solution is output, and an updated management strategy is generated accordingly. This mechanism realizes the dynamic response and continuous correction of the periodic optimization scheme during the seedling process, thereby improving the robustness and environmental adaptability of the system.

[0127] In this embodiment, the S7 specifically includes:

[0128] S71, receiving a seedling management execution strategy or an updated seedling management execution strategy, converting the stage task execution schedule T into a set of structured task scheduling entries, each entry encapsulating a task identifier, a target stage, an execution time, and a required resource category label in the form of a key-value pair, to form a time scheduling mapping set;

[0129] S72, mapping the resource allocation configuration table R to a set of allocation instructions, each allocation instruction is a structure triple<k,j,m> →r kjm Indicates the allocation quantity for the corresponding phase, task, and resource type;

[0130] S73, serialize and encode the management action instruction sequence set into an action control instruction stream, and generate a control instruction format set that matches the communication interface of each physical control node;

[0131] S74: The scheduling mapping set, the allocation instruction set, and the control instruction format set are aggregated and encapsulated into a policy synchronization object, and a policy deployment request is initiated to the seedling management platform to complete the policy push of the periodic scheduling result;

[0132] S75. The seedling management platform parses the strategy synchronization object and sends each task scheduling entry, resource allocation item and control instruction therein to each scheduling control sub-module in the platform to complete the system-level deployment of the seedling management execution strategy and task scheduling execution initialization.

[0133] The present invention forms a complete policy synchronization object by converting the seedling management execution strategy into a set of task scheduling entries, a set of resource allocation instructions, and a control instruction stream, thereby supporting deployment requests for the seedling management platform. Each task entry clearly encapsulates the task identifier, execution stage, time, and resource category in the form of a key-value pair. The resource instruction represents the resource allocation information in a structural triplet, and the control instruction is encoded in a communication format that can be recognized by the physical control node. After the platform receives the policy synchronization object, it can automatically parse and send it to the scheduling control module, realizing the rapid deployment and precise execution of the system-level seedling management strategy, thereby improving the scheduling efficiency and management automation level.

[0134] Example 1:

[0135] In order to verify the feasibility of the present invention in implementation, the present invention was applied to a large-scale container seedling management project in a forest nursery base. The base is located in a typical monsoon climate zone with diverse soil types and a seedling cycle of up to 150 days, involving multiple varieties of native tree species. Traditional management methods rely on planting experience to formulate fixed timetables and resource allocation plans, which cannot dynamically respond to environmental changes. Problems such as extended seedling cycles, waste of resources, and unstable seedling survival rates are prone to occur, especially under the background of continuous rainfall and large seasonal temperature differences. This embodiment aims to solve the problems of scheduling rigidity, feedback lag, and inaccurate health control through the forest seedling cycle management optimization method based on the adaptive algorithm proposed by the present invention.

[0136] During the implementation process, the project deployed environmental sensing nodes and seedling status sensing equipment to continuously collect five types of environmental parameters, including temperature, humidity, light intensity, soil moisture, and soil nutrients, as well as eight growth status indicators, including seedling height, growth rate, root development index, stem diameter, leaf area index, root-crown ratio, physiological maturity score, and disease and insect pest incidence, to construct an original indicator data set. Subsequently, the method of the present invention was used to clean and normalize the data, extract the standardized indicator matrix, and construct the seedling stage label in combination with seedling height, root development, and growth rate. By fusing the input features, a multi-objective periodic scheduling model was constructed, and three types of objective functions, including total seedling duration, resource consumption, and health score, were output. The optimal solution for periodic scheduling was generated based on the entropy weight adaptive gray wolf optimization algorithm, and converted into a task execution schedule, a resource allocation table, and a management action instruction set, which were then deployed in the management platform.

[0137] During the implementation period, real-time feedback data triggered six rescheduling operations, corresponding to unusual changes in conditions such as sudden rainfall, dramatic fluctuations in soil drying, stagnant seedling growth, and unusually high temperatures. After each deviation was automatically identified, the model reconfigured and updated its strategy within an average of eight seconds, allowing the system to automatically adjust task scheduling and resource allocation.

[0138] Compared with the traditional manual cycle management method, the system of the present invention shows obvious advantages in controlling the total cycle time, resource consumption and seedling health. As shown in the table below, in the 150-day seedling cycle, after adopting the system of the present invention, the average seedling cycle was shortened by 13.7%, the comprehensive resource consumption was reduced by 21.4%, and the seedling health score was improved by 18.9%, among which the average root development value was improved most significantly, indicating that the optimized strategy is effective in promoting the physiological maturity of seedlings. The results show that the scheduling optimization method proposed in the present invention not only has the scientific nature of data-driven, but also has the stability and adaptability of actual implementation, which effectively improves the overall efficiency of forest seedling management.

[0139] The following table is a "Comparative Data Table of Seedling Management Optimization Effects Based on Adaptive Algorithms", which lists the comparisons between traditional management methods and the method of the present invention on multiple key indicators:

[0140] Table 1 Comparative data of seedling management optimization effect based on adaptive algorithm

[0141]

[0142] As can be seen from the above table, the present invention has a significant optimization effect on multiple key performance indicators of forest seedling cycle management compared to traditional management methods. First, in terms of cycle execution efficiency, the average seedling cycle under the traditional management method is 150 days. After adopting the optimization strategy based on the adaptive algorithm proposed by the present invention, the seedling cycle is shortened to 129.4 days, with an overall shortening of 13.7%. This shows that the strategy solved by the multi-objective cycle scheduling model and the entropy weight adaptive gray wolf optimization algorithm can more effectively coordinate task scheduling and phased resource input, thereby improving the execution efficiency of the entire seedling process.

[0143] Secondly, in terms of resource usage, the present invention demonstrates superior regulatory capabilities over traditional management in the three dimensions of total water consumption, fertilizer consumption, and manpower input. Specifically, the total consumption of water and fertilizer was reduced to 78.6% and 81.2% of the traditional solution, respectively, and manpower input also dropped to 77.5%, with the overall resource savings exceeding 18%. This shows that the present invention can accurately assess the resource requirements of each stage during the resource allocation strategy generation phase, and avoid resource waste through structured execution strategies. In addition, in terms of the total duration of light utilization, the scheme of the present invention has increased the available light time by 5.8% through the reasonable optimization of the operation sequence in the seedling stage, effectively improving the efficiency of natural resource utilization.

[0144] Finally, the present invention also shows obvious advantages in terms of seedling growth health and system responsiveness. In terms of health scores, the average health score of seedlings increased from 76.2 under the traditional method to 90.6, an increase of 18.9%, of which the root development index increased by 27.7%, which fully demonstrates that the optimized management strategy helps seedlings form a more robust underground structure. The incidence of pests and diseases decreased from 12.1% to 6.3%, a decrease of 47.9%, reflecting that the dynamic scheduling strategy has stronger intervention and early warning capabilities for environmental anomalies and pest and disease risks. At the same time, the system can realize automatic strategy updates, with an average response time of only 8 seconds, and automatically triggers feedback corrections up to 6 times during the entire seedling cycle, which is far superior to the traditional manual passive response method, and significantly enhances the system's adaptability to environmental fluctuations and scheduling stability. Overall, the present invention has achieved systematic improvements in the three aspects of seedling cycle, resource input and seedling quality through optimized scheduling, dynamic response and structured execution strategies.

[0145] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A forest seedling cycle management optimization method based on an adaptive algorithm, characterized in that: The steps include: S1. Collect multi-source environmental data and seedling growth status data during forest seedling cultivation to generate a set of original indicator data; S2. Process the original indicator data set to obtain a standardized indicator matrix, and construct a stage indicator evaluation input set corresponding to the standardized indicator matrix in combination with the seedling stage labels; S3. Construct a multi-objective periodic scheduling model based on the phased indicator evaluation input set and output a multi-objective scheduling function set; S4. Introduce the entropy-weighted adaptive grey wolf optimization algorithm to solve the multi-objective scheduling function set. By calculating the information entropy of each indicator, the entropy weight vector is generated, the fusion objective function is constructed and iteratively updated, and the optimal solution for periodic scheduling is output. S5. Generate a seedling management execution strategy based on the optimal solution of periodic scheduling; S6. The environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, the feedback monitoring data set is compared with the original indicator data set, an updated periodic scheduling optimal solution is output, and an updated seedling management execution strategy is generated; S7. Synchronize the seedling management execution strategy or the updated seedling management execution strategy to the seedling management platform to complete the implementation deployment and task scheduling execution of the cycle optimization results.

2. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The original indicator data set includes environmental parameters and seedling growth status parameters; the environmental parameters include temperature, humidity, light intensity, soil moisture content and soil nutrient concentration; the seedling growth status parameters include seedling height, growth rate, stem diameter, leaf area index, root development index, root-to-crown ratio, physiological maturity score and disease and insect pest incidence.

3. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S2 specifically includes: S21. Perform data cleaning operations on the data of each dimension in the original indicator data set, remove missing values, duplicate values, and abnormal data exceeding the statistical threshold, and generate a cleaned indicator data set; S22. Perform normalization processing on each dimension indicator in the cleaned indicator data set, calculate the normalized result of each indicator value using a linear normalization method, arrange all normalized results by sample, and construct a normalized indicator matrix; S23. According to the normalized values ​​of seedling height, growth rate, and root development index corresponding to each sample in the standardized indicator matrix, a threshold judgment rule for seedling stage identification is set, and the stage to which each sample belongs is automatically determined based on the threshold judgment rule to generate a seedling stage label set; S24. Combine the standardized indicator matrix with the seedling stage label set to construct a stage indicator evaluation input set.

4. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S3 specifically includes: S31, constructing a multi-objective periodic scheduling model, wherein the multi-objective periodic scheduling model includes an indicator input fusion layer, a stage duration evaluation layer, a resource usage estimation layer, and a health score aggregation layer; S32. In the indicator input fusion layer, each standardized indicator vector in the stage indicator evaluation input set is input into the fusion layer, and dimension uniform mapping, feature reconstruction and normalization processing are performed to generate a fusion input feature set; S33. In the stage duration evaluation layer, based on the fused input feature set and the corresponding seedling stage label, the duration estimate of each seedling stage is calculated respectively, and the duration estimates of all seedling stages are accumulated to obtain the total seedling duration objective function; S34. In the resource usage estimation layer, based on the resource-related indicators in the fused input features, the resource usage of each type involved in each seedling stage is counted separately, and the resource usage of all resource types in all seedling stages is summed up to construct a resource consumption objective function; S35. In the health score aggregation layer, extract the three dimensions of seedling height, growth rate, and root development index from the fused input features, calculate the health score value of each sample separately, and sum the health scores of all samples in all seedling stages and divide them by the total number of samples to obtain the seedling health score objective function; S36. The total seedling raising time objective function, the resource consumption objective function and the seedling health score objective function are combined to form a multi-objective scheduling function set F = {f1, f2, f3}.

5. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S4 specifically includes: S41, initializing the search space dimension of the gray wolf optimization algorithm, determining the search dimension according to the input parameter structure of the multi-objective periodic scheduling function set, and randomly generating initial position vectors of multiple gray wolf individuals to form an initial population; S42. Calculate the fitness value of each individual gray wolf under the objective function of total seedling cultivation time, the objective function of resource consumption, and the objective function of seedling health score, and calculate the information entropy for each objective function based on the normalized fitness of all individuals under each objective function; S43. Calculate the entropy weight according to the information entropy corresponding to each objective function. The smaller the information entropy, the greater the weight of the objective function. The sum of the entropy weights of all objective functions is 1. S44, weighted multiplication of the entropy weight of each objective function and the corresponding individual fitness value, and summing the results over the three objective function dimensions to obtain the fusion fitness value corresponding to each gray wolf individual; S45. In each iteration, the gray wolf individuals are sorted according to the fusion fitness value, and the currently optimal, second-optimal, and third-optimal individuals are selected as guide individuals. The new positions of all gray wolf individuals are calculated according to the position update mechanism of the standard gray wolf optimization algorithm. S46. Repeat the fitness calculation, entropy weight update and position update processes until the preset maximum number of iterations is reached, and finally output the individual position vector with the optimal fusion fitness value. The parameter solution corresponding to the optimal individual position vector is the optimal solution for periodic scheduling.

6. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 5, characterized in that: The information entropy calculation includes normalizing the fitness value of each gray wolf individual under multiple objective functions, constructing a normalized fitness sequence for each objective function, and calculating the information entropy of the j-th objective function according to the information entropy formula: Among them, e j represents the information entropy value of the jth objective function in the multi-objective scheduling function set, p ij represents the normalized fitness value of the i-th gray wolf individual under the j-th objective function, n represents the number of individuals in the gray wolf population, k represents the normalization constant, ε represents a positive constant to prevent the logarithmic term from being zero, i represents the index number of the gray wolf population individual, and j represents the index number of the objective function in the multi-objective scheduling function set.

7. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 5, characterized in that: The S45 specifically includes: S451, the position update mechanism of the standard gray wolf optimization algorithm includes guiding individual selection, prey position estimation, guiding vector construction and position fusion update; S452, guide individual selection: all gray wolf individuals are ranked based on the fusion fitness value, and the top three individuals with the best fitness values ​​are selected, which are defined as guide individuals α, β and δ respectively; S453. Prey Position Estimation: For any individual i in the gray wolf population, calculate the directional distance vector of any individual i relative to the guide individual α, β, and δ, and generate the perturbation coefficient and position offset based on the current search round to estimate the target prey position of any individual i in the three directions; S454: Execute a guidance position vector construction operation, and construct a fitted position update expression for each direction guidance result based on the three estimated positions and the corresponding distance control vectors; S455. Perform a position fusion update operation, perform a mean fusion operation on the position vectors obtained by guiding in the three directions, and obtain the new position vector of the i-th gray wolf individual in this round of iteration as the update result of the current search state.

8. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S5 specifically includes: S51. Express the optimal solution of periodic scheduling as a structure vector where s k represents the scheduling time and resource allocation results of the kth seedling raising stage, where k is the total number of seedling raising stages; S52, according to the structure vector The scheduling time of each stage in the construction phase task execution schedule is a two-dimensional array T = {t kj }, where t kj represents the planned execution time of the jth task in the kth stage; S53, according to the structure vector The resource consumption components of each stage in the process are used to construct a resource allocation configuration table, which is a three-dimensional structure R={r kjm }, where r kjm It represents the amount of resources of type m allocated to type j management action in the kth stage. The resource types include water, fertilizer, light and manpower. S54, generating a management action instruction sequence for each stage based on the stage task execution schedule T and the resource allocation configuration table R, wherein the instruction sequence is a management action instruction sequence set; S55. Integrate the task execution schedule T, the resource allocation configuration table R and the management action instruction sequence set to form a complete seedling management execution strategy.

9. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S6 specifically includes: S61. During the implementation of the seedling management execution strategy, the environmental feedback data and seedling feedback data collected in real time are combined into a feedback monitoring data set, and feature dimension alignment and time alignment are performed with the original indicator data set; S62, performing deviation calculation on the key indicators of the same dimension in the feedback monitoring data set and the original indicator data set to obtain an indicator deviation vector Δ(t); S63, compare all the index items in the index deviation vector Δ(t) one by one with the preset threshold vector θ=[θ1,θ2,…,θ i ,…,θ m ] to compare, if there is any one that satisfies Δ (t) ≥θ i , it is determined that the current feedback monitoring data has triggered an abnormal deviation; S64. When the deviation judgment result is a trigger state, the multi-objective periodic scheduling model is reconstructed using the phased indicator evaluation input set and the entropy weight adaptive grey wolf optimization algorithm, the updated periodic scheduling optimal solution is output, and the updated seedling management execution strategy is generated.

10. The method for optimizing forest seedling cycle management based on an adaptive algorithm according to claim 1, characterized in that: The S7 specifically includes: S71, receiving a seedling management execution strategy or an updated seedling management execution strategy, converting the stage task execution schedule T into a set of structured task scheduling entries, each entry encapsulating a task identifier, a target stage, an execution time, and a required resource category label in the form of a key-value pair, to form a time scheduling mapping set; S72, mapping the resource allocation configuration table R to a set of allocation instructions, each allocation instruction is a structure triple<k,j,m> →r kjm Indicates the allocation quantity for the corresponding phase, task, and resource type; S73, serialize and encode the management action instruction sequence set into an action control instruction stream, and generate a control instruction format set that matches the communication interface of each physical control node; S74: The scheduling mapping set, the allocation instruction set, and the control instruction format set are aggregated and encapsulated into a policy synchronization object, and a policy deployment request is initiated to the seedling management platform to complete the policy push of the periodic scheduling result; S75. The seedling management platform parses the strategy synchronization object and sends each task scheduling entry, resource allocation item and control instruction therein to each scheduling control sub-module in the platform to complete the system-level deployment of the seedling management execution strategy and task scheduling execution initialization.

Images