Natural reserve resource multi-objective optimization method and device, storage medium and equipment

Abstract

The present application relates to a nature reserve resource multi-objective optimization method, device, storage medium and equipment, by acquiring the multi-source remote sensing data of the target nature reserve, using the pre-trained prediction model to realize the prediction of the aboveground biomass data and plant crude protein data of the target nature reserve, and based on the second generation non-dominated sorting genetic algorithm, according to the biological data, aboveground biomass data, plant crude protein data, decision target and constraint condition of the target nature reserve, the best planning information of the target nature reserve is obtained, which provides data support for the planning of the livestock structure, wild animal quantity and artificial grassland planting information of the target nature reserve, solves the balance problem of multiple targets of human settlement-grassland-wild animal-livestock balance in nature reserve, and promotes the sustainable development of nature reserve.

Description

Technical Field

[0001] This invention relates to the field of land use planning, and in particular to a method, apparatus, storage medium, and equipment for multi-objective optimization of resources in nature reserves. Background Technology

[0002] Nature reserves play a vital role in protecting alpine ecosystems and safeguarding national ecological security. However, due to the impacts of climate change and human activities, nature reserves face challenges such as malnutrition in wild herbivores and livestock, degradation of grassland biodiversity and multifunctionality, and reduced income for herders.

[0003] Previous studies on the resource utilization of nature reserves have mainly focused on the supply of grassland forage and the demand of livestock for forage, which has a single focus and makes it difficult to achieve precise allocation and sustainable utilization of resources in nature reserves. Summary of the Invention

[0004] This application provides a method, apparatus, storage medium, and device for multi-objective optimization of resources in nature reserves, which can provide data support for planning resources such as livestock structure, wildlife population, and artificial grassland planting in target nature reserves, thereby promoting the sustainable development of nature reserves.

[0005] In a first aspect, embodiments of this application provide a multi-objective optimization method for resources in nature reserves, including:

[0006] Acquire biological data and multi-source remote sensing data of the target nature reserve; wherein the biological data includes grassland data, livestock data and wildlife data;

[0007] Based on the multi-source remote sensing data and the pre-trained prediction model, obtain aboveground biomass data and plant crude protein data of the target nature reserve;

[0008] Obtain the decision objectives and constraints of the target nature reserve; wherein, there are at least two decision objectives, and the decision objectives and constraints are related to at least one of the aboveground biomass data, crude protein data, grassland data, and animal data;

[0009] Based on the second-generation non-dominated sorting genetic algorithm, the optimal planning information of the target nature reserve is obtained according to the biological data, aboveground biomass data, plant crude protein data, decision objective and constraints of the target nature reserve; wherein, the optimal planning information includes livestock structure, wild animal population and artificial grassland planting information.

[0010] Secondly, embodiments of this application provide a multi-objective optimization device for resources in nature reserves, comprising:

[0011] The data acquisition module is used to acquire biological data and multi-source remote sensing data of the target nature reserve; wherein, the biological data includes grassland data, livestock data and wildlife data;

[0012] The data prediction module is used to obtain aboveground biomass data and plant crude protein data of the target nature reserve based on the multi-source remote sensing data and the pre-trained prediction model.

[0013] The decision information acquisition module is used to acquire the decision objectives and constraints of the target nature reserve; wherein, there are at least two decision objectives, and the decision objectives and constraints are related to at least one of the aboveground biomass data, crude protein data, grassland data and animal data;

[0014] The planning information acquisition module is used to acquire the optimal planning information of the target nature reserve based on the second-generation non-dominated sorting genetic algorithm, according to the biological data, aboveground biomass data, plant crude protein data, decision objectives, and constraints of the target nature reserve; wherein, the optimal planning information includes livestock structure, wild animal numbers, and artificial grassland planting information.

[0015] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the multi-objective optimization method for resources in nature reserves as described in any of the preceding claims.

[0016] Fourthly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable by the processor;

[0017] When the processor executes the computer program, it implements the steps of the multi-objective optimization method for resources in nature reserves as described in any of the above.

[0018] In this embodiment, multi-source remote sensing data of the target nature reserve is acquired, and a pre-trained prediction model is used to predict the aboveground biomass data and plant crude protein data of the target nature reserve. Then, based on the second-generation non-dominated sorting genetic algorithm, the optimal planning information of the target nature reserve is obtained according to the biological data, aboveground biomass data, plant crude protein data, decision objectives and constraints of the target nature reserve. This provides data support for the planning of livestock structure, wildlife population and artificial grassland planting information of the target nature reserve, solves the balance problem of multiple objectives of human settlement-grassland-wildlife-livestock balance in nature reserves, and promotes the sustainable development of nature reserves.

[0019] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a flowchart of a multi-objective optimization method for resources in nature reserves according to one embodiment of the present invention;

[0021] Figure 2 This is a flowchart of a multi-objective optimization method for resources in nature reserves, as described in another embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of step S120 in one embodiment of the present invention;

[0023] Figure 4 This is a spatial distribution map of aboveground biomass (AGB) and crude protein (CP) in the Qiangtang Nature Reserve, as predicted in one embodiment of the present invention.

[0024] Figure 5 This is a spatiotemporal distribution map of the seasonal imbalance between aboveground biomass supply to livestock in different areas of the Qiangtang Nature Reserve, as shown in one embodiment of the present invention.

[0025] Figure 6 This is a spatiotemporal distribution map of the seasonal imbalance between crude protein supply to livestock in the Qiangtang protected area, as described in one embodiment of the present invention.

[0026] Figure 7 This is a schematic diagram of a multi-objective optimization device for resources in a nature reserve according to one embodiment of the present invention;

[0027] Figure 8 This is a schematic diagram of the structure of a computer device according to one embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0029] It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.

[0030] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0031] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0032] Furthermore, in the description of this application, unless otherwise stated, "several" refers to two or more. "And / or" describes the correspondence between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0033] Grassland-livestock balance refers to maintaining a reasonable carrying capacity of grasslands in the long term and making rational use of grasslands to maintain the healthy cycle of the grassland ecosystem. In nature reserves, grazing must consider not only the grass-livestock balance but also the protection of wildlife. The concept of grass-livestock balance has gradually evolved into a balance problem of "human settlement-grassland-wildlife-livestock," with the goal of meeting the living standards of herders, the ecological environment of grasslands, and the nutritional needs of wildlife and livestock. The spatial mismatch of pasture resources, competition for distribution space between wildlife and livestock, herders' income levels, and the protection of grassland ecosystems make solving the "human settlement-grassland-wildlife-livestock" balance a key factor in achieving the sustainable development of nature reserves.

[0034] Previous studies on grassland-livestock balance have mainly focused on the supply of grassland forage and the demand of livestock for forage. For example, Ma et al. (2024) used the CASA model to estimate grassland net primary productivity (NPP) and then converted it into aboveground biomass (AGB) to assess the grassland-livestock balance in Xinjiang from 1982 to 2020. Umuhoza et al. (2021) used vegetation indices (such as MOD1 products) to convert NPP data into aboveground biomass, estimated the productivity of mountain grasslands in Kyrgyzstan and Tajikistan, and assessed their carrying capacity.

[0035] However, these studies focus on a single objective and fail to achieve a balance between multiple objectives in nature reserves, such as human settlements, grasslands, wild herbivores, and livestock.

[0036] Therefore, for the above issues, please refer to [link / reference]. Figure 1 This application provides a multi-objective optimization method for resources in nature reserves, the method comprising:

[0037] S110: Acquire biological data and multi-source remote sensing data of the target nature reserve; wherein, the biological data includes grassland data, livestock data and wildlife data;

[0038] Optionally, the grassland data may include data such as the area of ​​natural grassland and the area of ​​artificial grassland.

[0039] The animal data may include data such as the number of livestock in stock, the number of livestock slaughtered, the amount of above-ground food consumed by livestock, the number of wild animals, and the daily food consumption of wild animals;

[0040] S120: Based on the multi-source remote sensing data and the pre-trained prediction model, obtain aboveground biomass data and plant crude protein data of the target nature reserve;

[0041] Predictive models can be pre-built and trained by users to predict aboveground biomass data and plant crude protein data.

[0042] Predictive models can be built based on deep neural network algorithms or other deep learning algorithms.

[0043] Deep learning is a machine learning technique based on artificial neural networks that models complex data through multi-level nonlinear transformations. Compared to traditional machine learning methods, deep learning can automatically learn feature extraction and performs exceptionally well when processing high-dimensional, nonlinear data. The model typically consists of multiple neural network layers (such as convolutional neural networks and recurrent neural networks), optimizing parameters through forward and backward propagation.

[0044] S130: Obtain the decision objectives and constraints of the target nature reserve; wherein, there are at least two decision objectives, and the decision objectives and constraints are related to at least one of the aboveground biomass data, crude protein data, grassland data, and animal data;

[0045] Decision-making objectives and constraints can be set based on the characteristics of grasslands and animal husbandry in the target nature reserve.

[0046] The decision objective is the optimization objective of the second-generation non-dominated sorting genetic algorithm. The second-generation non-dominated sorting genetic algorithm can simultaneously maximize or minimize multiple conflicting decision objectives.

[0047] Decision-making objectives can include economic decision-making objectives, productivity decision-making objectives, and ecological and environmental decision-making objectives.

[0048] Constraints are used to constrain individuals in the second-generation non-dominated sorting genetic algorithm. In this embodiment, constraints can be used to constrain the area of ​​natural grassland and artificial grassland, and / or constraints can be used to constrain crude protein.

[0049] S140: Based on the second-generation non-dominated sorting genetic algorithm, the optimal planning information of the target nature reserve is obtained according to the biological data of the target nature reserve, the aboveground biomass data, the plant crude protein data, the decision objective, and the constraints; wherein, the optimal planning information includes livestock structure, wild animal numbers, and artificial grassland planting information.

[0050] The second-generation non-dominated sorting genetic algorithm (NSGA-II) is an improved version of the non-dominated sorting genetic algorithm. Based on the basic genetic algorithm, it improves the selection and regeneration method by stratifying each individual according to their dominance and non-dominance relationships before performing selection operations. This results in highly satisfactory results in multi-objective optimization. The basic algorithm flow is as follows: First, an initial population is generated, typically a randomly generated population of size N. This population is then sorted using non-dominance methods, and the first generation of offspring is obtained through the three basic operations of the genetic algorithm: selection, crossover, and mutation. Then, starting from the second generation, the offspring population is merged with the parent population. A fast non-dominated sort is performed, and the crowding degree of individuals in each non-dominated layer is calculated. Based on the individuals' non-dominance relationships and crowding degrees, N individuals are selected according to selection criteria to form a new next-generation population. Finally, selection, crossover, and mutation operations are used again to generate a new offspring population, and this cycle continues until the program terminates.

[0051] In this embodiment, multi-source remote sensing data of the target nature reserve is acquired, and a pre-trained prediction model is used to predict the aboveground biomass data and plant crude protein data of the target nature reserve. Then, based on the second-generation non-dominated sorting genetic algorithm, the optimal planning information of the target nature reserve is obtained according to the biological data, aboveground biomass data, plant crude protein data, decision objectives and constraints of the target nature reserve. This provides data support for the planning of livestock structure, wildlife population and artificial grassland planting information of the target nature reserve, solves the balance problem of multiple objectives of human settlement-grassland-wildlife-livestock balance in nature reserves, and promotes the sustainable development of nature reserves.

[0052] like Figure 2 As shown, in one embodiment, multi-source remote sensing data may include data such as remote sensing spectra, topographic indices, vegetation indices, and soil physicochemical parameters.

[0053] After acquiring multi-source remote sensing data, the following is included:

[0054] S210: Based on the Pearson correlation analysis algorithm, select relevant variables that are correlated with aboveground biomass and plant crude protein from the at least two variables;

[0055] Pearson correlation analysis is a method used to measure the linear relationship between two variables. It determines the correlation between two variables by calculating the Pearson correlation coefficient, which is the quotient of the product of the covariance and standard deviation of the two variables.

[0056] S220: Based on the Boruta feature selection algorithm, select significant variables that are significantly correlated with the aboveground biomass and the plant crude protein from the relevant variables to obtain a feature subset;

[0057] The Boruta feature selection algorithm constructs shadow variables as a benchmark for feature importance and combines them with a random forest model for multiple rounds of importance comparison and statistical testing. Specifically, the system first generates random perturbation shadow variables with the same dimensions as the original variables, incorporates them into the feature space, trains the random forest model, and calculates the importance scores of each real variable and its shadow variable. Then, it uses hypothesis testing methods (such as T-tests) to identify core features with significantly higher importance than the optimal shadow variable, while eliminating insignificant variables with importance below a threshold. Through multiple iterations, the feature set is continuously refined until all variables are clearly categorized as either retained or eliminated, ultimately selecting a subset of statistically significant strongly correlated features. This process effectively balances the robustness and accuracy of feature selection through dynamically adjusting the significance threshold and a full-round validation mechanism.

[0058] The selected feature subset may include multiple environmental factors that are significantly associated with aboveground biomass and plant crude protein.

[0059] As shown in Table 1, the feature subset may include climate-related environmental factors such as annual mean temperature, annual precipitation, annual mean evapotranspiration, and annual mean photosynthetically active radiation; vegetation-related environmental factors such as normalized vegetation index, difference vegetation index, ratio vegetation index, conversion vegetation index, atmospheric corrected ratio vegetation index, leaf area index, and net primary productivity; and topographic-related environmental factors such as altitude, slope, aspect, and topographic humidity index.

[0060] Table 1 Feature Subset

[0061]

[0062] S230: Obtain a dataset based on the feature subset, and use the dataset to pre-train a deep neural network model to obtain a prediction model.

[0063] After determining the feature subset, specific data for each significant variable in the feature subset can be obtained from multi-source remote sensing data of the target nature reserve.

[0064] Multi-source remote sensing data can include data from remote sensing datasets such as the GEE ERA5-Land dataset, the GEE MOD15A2H dataset, the GEEMOD16A2 dataset, the GEE USGS dataset, and the Calculated from ASTER GDEM dataset.

[0065] Among them, the annual mean temperature and annual precipitation can be obtained from the GEE ERA5-Land dataset, the annual mean evapotranspiration can be obtained from the GEEMOD15A2H dataset, the annual mean photosynthetically active radiation can be obtained from the GEE MOD16A2 dataset, the vegetation-related normalized vegetation index, difference vegetation index, ratio vegetation index, conversion vegetation index, atmospheric corrected ratio vegetation index, leaf area index, and net primary productivity can be obtained from the GEE USGS dataset, and the topographic-related altitude, slope, aspect, and topographic humidity index can be obtained from the Calculated from ASTER GDEM dataset.

[0066] Deep neural network models can be built using TensorFlow and Keras frameworks. Deep neural network models optimize network parameters through forward and backward propagation, combine Dropout layers and early stopping to suppress overfitting, and perform cross-validation by dividing a validation set during training to finally obtain the optimal model weights.

[0067] The dataset in this application contains data on significant variables that are significantly associated with aboveground biomass and plant crude protein. Using this dataset to pre-train a deep neural network model can reduce the amount of data processing and improve pre-training efficiency.

[0068] Specifically, the dataset is divided into training, validation, and test sets using a random sampling method. The network structure of the deep neural network model is selected based on the data, and the model parameters are initialized. The model hyperparameters (such as learning rate, batch size, number of layers, etc.) are optimized by grid search combined with cross-validation. The performance of different hyperparameter combinations is evaluated using the validation set, and the optimal parameters are selected.

[0069] Using goodness of fit ( The accuracy of model predictions is comprehensively evaluated using indicators such as root mean square error (RMSE) and mean absolute error (MAE). Reflecting the explanatory power of independent variables on the dependent variable, RMSE and MAE measure the dispersion and absolute deviation between predicted and measured values, respectively. The formulas for the three model evaluation indicators are as follows:

[0070]

[0071]

[0072]

[0073] Where n is the number of samples, Here is the observed aboveground biomass value for sample i. Here is the predicted aboveground biomass value for sample i. This represents the average aboveground biomass of all samples.

[0074] In this embodiment, the correlation of all variables in the multi-source remote sensing data is preliminarily analyzed using the Pearson correlation analysis algorithm to identify the relevant variables that are correlated with aboveground biomass and plant crude protein. Then, the Boruta feature selection algorithm is used to screen out the significant variables with significant correlations to generate a feature subset. Based on the feature subset, a dataset that is significantly correlated with aboveground biomass and plant crude protein is obtained. By pre-training the deep neural network model, the amount of data processing is reduced and the prediction efficiency is improved.

[0075] like Figure 3 As shown, in one embodiment, based on the multi-source remote sensing data and the pre-trained prediction model, aboveground biomass data and plant crude protein data of the target nature reserve are obtained, including:

[0076] S121: Preprocess the multi-source remote sensing data; wherein, the preprocessing includes missing value handling, feature standardization, and spatial alignment;

[0077] Multi-source remote sensing images can include remote sensing images from different data sources, different distributions, or different dimensions.

[0078] For missing values ​​in multi-source remote sensing images, spatial interpolation, spatiotemporal interpolation, and other methods can be used to fill in the missing values.

[0079] Feature standardization is used to adjust features of different scales to the same scale. Existing feature standardization methods such as Z-Score and Max-Min can be used to achieve feature standardization.

[0080] Spatial alignment is used to project multi-source remote sensing images from different data sources, distributions, or dimensions into the same "comparable" mathematical space, thereby facilitating the use of subsequent prediction models. Specifically, spatial alignment of multi-source remote sensing data can be achieved through existing spatial alignment algorithms such as resampling, subspace mapping, and deep learning alignment.

[0081] S122: Perform unified resampling and coordinate transformation on the spatially heterogeneous raster data in the multi-source remote sensing data, and unify the raster data to the target resolution;

[0082] Spatial heterogeneity refers to the uneven distribution of ecological processes and patterns across different geographical locations.

[0083] Optionally, in this embodiment, the rasterio library is used to perform unified resampling and coordinate transformation on multi-resolution remote sensing images, and the reproject function is used to unify raster data from different sources to a resolution of 256×256 pixel blocks to ensure the matching of input features in spatial scale.

[0084] S123: Randomly select sampling points in the target nature reserve, obtain aboveground biomass and plant crude protein data at the sampling points, and obtain sample data;

[0085] The number and size of sampling points can be set according to actual needs. In this embodiment, 150 sampling points are randomly selected in the target nature reserve. Three 1 m × 1 m quadrats are randomly set up at each sampling point, and the aboveground biomass (AGB) harvested from each 0.5 m × 0.5 m quadrats is used as a representative sample. Topsoil (0-30 cm) is collected using a five-point sampling method. The five subsamples are mixed and placed in sealed plastic bags for laboratory analysis. Three replicates are performed for each plot, and the latitude and longitude coordinates of each sampling point are recorded. Aboveground biomass is determined by uniformly harvesting herbaceous plants within 0.5 m × 0.5 m quadrats, drying them in a 65℃ constant temperature oven to constant weight, and then converting the dry weight to g / m² as the grassland aboveground biomass.

[0086] Crude protein (CP) from plants was determined using an automated elemental analyzer (Elementar vario MACROcube, Germany).

[0087] S124: Generate a dataset based on the feature subset and sample data, and use the dataset to pre-train the deep neural network model to obtain a prediction model.

[0088] The dataset includes feature subsets of data and sample data. The feature subsets of data can be obtained from existing remote sensing image databases, websites, or datasets.

[0089] Specifically, the existing process_single_band_rasters function can be used to read each band of data block by block, input into the trained prediction model for parallel computation, generate pixel-by-pixel prediction values, and after obtaining the prediction values, output the spatial distribution raster layer in GeoTIFF format.

[0090] Preferably, to enhance the regional applicability of the mapping results, the rasterio.mask module can be used to spatially crop the predicted raster based on the protected area boundary vector data (Shapefile), removing irrelevant areas and retaining the target range.

[0091] In this embodiment, by performing preprocessing such as missing value processing, feature standardization, and spatial alignment on multi-source remote sensing data, the spatially heterogeneous raster data in the multi-source remote sensing data is uniformly resampled and coordinate transformed to unify the raster data to the target resolution. Then, by reading several single-band raster data block by block, the efficiency of data processing and the accuracy of prediction can be improved when generating pixel-by-pixel aboveground biomass prediction values ​​and plant crude protein prediction values ​​using a pre-trained prediction model.

[0092] In one embodiment, animal data includes livestock data such as sheep, goats, and yaks, and wildlife data such as Tibetan wild asses, Tibetan antelopes, and wild yaks.

[0093] Livestock data can include data such as the slaughter rate of various livestock, the number of livestock sold during the year, the number of livestock at the end of the year, the quantity of livestock, and the amount of feed consumed by above-ground organisms.

[0094] Wildlife data can include information such as the population of various wild animals.

[0095] Plant data includes the area of ​​naturally available grassland, the area of ​​planted artificial grassland, and parameters of the planted plants.

[0096] The following uses the Qiangtang Nature Reserve as an example to illustrate the specific details of the proposed scheme:

[0097] The Qiangtang Nature Reserve (32°10′~36°32′N, 79°42′~92°05′E) is located in the heart of the Qinghai-Tibet Plateau, covering a total area of ​​298,000 square kilometers with an average altitude of 5,000 meters. The Qiangtang region exhibits distinct seasonal characteristics, with a long and harsh cold season and a short and mild warm season, resulting in significant seasonal differences. The average annual temperature is mostly below 0°C, ranging from -1 to -18°C, reaching -6°C in the northwest. The average annual precipitation is 50–300 mm, with over 80% concentrated between June and September. Annual sunshine hours range from 2,800 to 3,400 hours, and the annual total solar radiation exceeds 836 kilojoules per square centimeter, far exceeding that of other regions at the same latitude. However, the plateau's surface reflectivity is as high as 40%, limiting the actual solar radiation energy received by the ground. The reserve contains lakes with a total area exceeding 25,000 km², accounting for 25% of China's total lake area, making it the plateau lake region with the most lakes and the highest lake surface area in the world. The vegetation type is mainly alpine desert steppe, with obvious vertical zonation. Southern Qiangtang (4200-5000m) belongs to the subarctic zone, with about 20,000 herders and more than 1 million livestock in the reserve; Northern Qiangtang (>5000m) belongs to the frigid zone, with the core area and northern buffer zone remaining uninhabited and rich in wildlife resources; the unique ecological environment provides an ideal natural laboratory for studying the balance pattern of "human-grassland-wildlife-livestock" in the plateau ecosystem.

[0098] Natural grasslands are the primary food source for grazing livestock in the Qiangtang Nature Reserve, and their carrying capacity is a decisive factor in the scale of livestock production. The Qiangtang National Park encompasses only the Qiangtang National Nature Reserve, covering an area of ​​297,100 square kilometers. Grassland is the dominant land use type, covering 250,800 square kilometers, accounting for 84.3% of the national park's total area. Currently, the Qiangtang Nature Reserve has 9,055,100 hectares of usable grassland, of which 5,556,200 hectares are contracted grassland, representing 61.36% of the total usable grassland area. Field surveys in pastoral areas revealed that the grazing grassland area is approximately 50% of the total usable grassland area during both the warm and cold seasons. During grazing, policy requires the preservation of a portion of aboveground biomass for natural ecological restoration of the grassland. In this study, 50% of the peak aboveground biomass was preserved in both the warm and cold seasons.

[0099] Table 2. AGB Fresh Weight of Natural Grassland in Different Grazing Seasons

[0100]

[0101] The Qiangtang National Park encompasses only the Qiangtang National Nature Reserve, covering an area of ​​297,100 square kilometers. The reserve is divided into a core protected area and a general control area. The core protected area covers 220,800 square kilometers, accounting for 74.2% of the total national park area. The general control area, covering 76,700 square kilometers, includes traditional pastoral areas, key areas for ecological restoration, areas with concentrated infrastructure construction, and areas providing opportunities for public interaction and experience with nature. An outer support area, approximately one-quarter the size of the Qiangtang Reserve (about 60,000 square kilometers), was added to the study to support ecological protection, restoration, and multifunctional enhancement of the Qinghai-Tibet Plateau nature reserve. This outer support area is rich in water and heat resources, and artificial grasslands are primarily planted there. Their main function is to provide supplementary hay for livestock in the general control area during the cool season. Through screening for forage crops, six high-yield crops were selected: sweet sorghum, silage corn, triticale, oats, and arrowhead peas. Table 3 lists the planting parameters for the artificial grasslands.

[0102] Table 3 Planting Parameters for Artificial Grassland

[0103]

[0104] Grazing livestock converts pasture and feed into livestock products. Different grazing livestock have different production performances and exert varying degrees of stress on grasslands. Adjusting the livestock structure is crucial for reducing grassland grazing stress and increasing economic returns. In addition to adjusting the livestock structure, adjusting the herd structure, appropriately increasing the proportion of female livestock, and improving slaughter efficiency can also improve grassland utilization and increase livestock product income.

[0105] The main herbivorous livestock in the Qiangtang Nature Reserve include yaks, sheep, goats, and a small number of horses. Livestock products mainly consist of beef and mutton, wool, and dairy products. Due to difficulties in data collection, the number of livestock slaughtered is calculated by multiplying the number of livestock in stock by a slaughter rate of 33.4% (the average slaughter rate of the five counties in previous years). Table 4 lists the changes in the number of livestock in stock in each county and city of the Qiangtang Nature Reserve in recent years.

[0106] Table 4. Changes in the number of various types of grazing livestock in the Qiangtang Nature Reserve

[0107]

[0108] Table 5 lists the daily nutritional requirements of the main grazing livestock in the Qiangtang Nature Reserve.

[0109] Table 5 Daily nutritional requirements for different livestock

[0110]

[0111] Among them, the daily dry matter and crude protein requirements of livestock are as follows: the data for sheep and goats are from "Feed Nutrition Magazine (11~16) - Qiu Wenshi, 1986 Inaugural Issue (5)"; the data for yaks are from "Qinghai Provincial Local Standard: Nutritional Requirements of Yaks".

[0112] Table 6 lists the population sizes and daily nutrient intake indicators of major wild animals in the Qiangtang Nature Reserve of the Qinghai-Tibet Plateau during the warm and cold seasons. Significant differences exist in nutrient requirements among different species, with wild yaks having the highest daily nutrient intake and Tibetan antelopes the lowest, reflecting niche differentiation in body size and metabolic needs.

[0113] Table 6. Population and Nutritional Requirements of Wild Animals

[0114]

[0115] Planning information can include multiple indicators related to livestock species, herd structure, number of wild animals, and artificial grassland planting. These indicators may include annual livestock sales, number of wild animals, number of livestock in stock, amount of hay purchased, total area of ​​natural usable grassland, area of ​​artificial grassland, amount of supplementary feed required for aboveground biomass zones, and amount of supplementary feed required for crude protein zones.

[0116] In this application, MATLAB software can be used to execute multi-objective optimization algorithms, improving data processing efficiency. MATLAB is an advanced technical computing and visualization software whose main functions include numerical computation, data analysis, signal processing, image processing, control system design, simulation, optimization, and machine learning. MATLAB's underlying architecture is based on matrix computation, making it highly efficient in handling matrices and arrays. MATLAB possesses a large number of toolboxes and function libraries, including signal processing, control systems, image processing, and statistics toolboxes, meeting diverse application needs. These toolboxes cover various fields such as mathematics, statistics, simulation, electronics, bioinformatics, finance, and testing. MATLAB also supports multiple programming languages, including MATLAB scripting, MATLAB functional languages, and languages ​​such as C, C++, and Java, allowing for various customizations and extensions. This flexibility makes MATLAB an important tool in scientific research and engineering applications.

[0117] As shown in Table 7, the indicators set according to the characteristics of grassland and animal husbandry in the Qiangtang Nature Reserve in this embodiment of the application include the following indicators:

[0118] Table 7 Key Indicators of the "Human-Grassland-Wildlife-Livestock" Multi-Objective Programming

[0119]

[0120] In one embodiment, the decision objective may include any one or more of the following: livestock productivity optimization objective, economic benefit optimization objective, grassland ecological environment optimization objective, and grassland habitat quality objective.

[0121] In this embodiment of the application, the decision objectives include livestock productivity optimization objectives, economic benefit optimization objectives, grassland ecological environment optimization objectives, and grassland habitat quality objectives.

[0122] Among them, livestock productivity is related to herd structure and turnover. Adjusting herd structure and enhancing turnover are the main ways to improve livestock productivity. This application mainly considers increasing the adjustment range of herd structure, accelerating turnover speed, and increasing slaughter rate under the premise of grass-livestock balance.

[0123] Specifically, the objectives for optimizing livestock productivity include increasing the slaughter rate:

[0124]

[0125] in, For slaughter rate, This represents the annual sales volume of livestock i. The number of livestock j at the end of the year; in this embodiment, livestock includes sheep, goats, and yaks.

[0126] The corresponding MATLAB expression can be: Where A1 is the slaughter rate.

[0127] The economic benefit optimization objective is the economic goal in the "human-grass-wildlife-livestock" balance plan, with the main purpose of maximizing livestock output. In the calculation, the average annual output value per unit of livestock product for each type of livestock is calculated separately to determine the total output value of the main grazing livestock.

[0128] Specifically, the economic benefit optimization objective includes maximizing livestock output:

[0129]

[0130] in, Let p represent the average annual output value of meat, wool, and cashmere per sheep, q represent the annual output value per goat, and r represent the annual output value per yak. For the number of sheep, For the number of goats, Let t represent the number of yaks, t represent the average price (in yuan) of 1 kg of hay purchased from outside sources, and h represent the quantity of hay purchased (in kg). The area of ​​each planted artificial grassland ( ), For planting 1 Cost (yuan);

[0131] The corresponding MATLAB expression can be: A2 = 2000 * x(1) + 2000 * x(2) + 20000 * x(3)) – x(10) – 3000*x(12) - 3879*x(13) - 2632*x(14) - 3000*x(15) - 1053*x(16), where A2 is the value of livestock production.

[0132] The goal of grassland ecological environment optimization is the ecological environment objective in the "human-grass-wildlife-livestock" balance plan, which can be achieved by minimizing animal grazing, i.e. minimizing grassland resource pressure.

[0133] Specifically, the goals for optimizing the grassland ecological environment include minimizing pressure on grassland resources:

[0134]

[0135] in, Pressure on grassland resources This represents the year-end inventory of each type of livestock. For each livestock, the amount of above-ground organisms consumed; For the number of each wild animal, This refers to the daily food intake of various wild animals. The number of grazing days in natural grassland of type k (k=1 represents the warm season, which is 153 days; k=2 represents the cold season, which is 212 days).

[0136] The corresponding MATLAB expression can be: A3=153* (x(7) * 5 + x(8) * 4.5 + x(9) * 30 + x(4) * 20 + x(5) * 7 + x(6) * 30) + 212*( x(7) * 5 + x(8) * 4.5 + x(9) * 30 + x(4) * 25 + x(5) * 10 + x(6) * 35). Where A3 represents grassland resource pressure.

[0137] Grassland habitat quality is a habitat quality target in the "human-grass-wildlife-livestock" balance planning. The habitat quality target aims to achieve ecosystem balance and stability by optimizing the aboveground biomass of natural and artificial grasslands.

[0138] Specifically, habitat quality is obtained in the following ways:

[0139]

[0140] in, The habitat quality is represented by u, the total usable area of ​​natural grassland, and d, the mean aboveground biomass of the protected area. This represents the area of ​​each planted artificial grassland. This represents the yield per acre of each artificial grassland.

[0141] The corresponding MATLAB expression can be: A4 = x(9)*d + 175506*x(12) + 38070*x(13) +33847*x(14) + 45524*x(15) + 21455*x(16), where A4 is the habitat quality.

[0142] Constraints may include livestock production constraints and grass-livestock balance constraints;

[0143] Livestock production constraints are used to keep the slaughter rate of livestock within a set range:

[0144]

[0145] in, and These are the upper and lower limits of the slaughter rate, which can be set as appropriate constants based on the slaughter rates of various counties in previous years.

[0146] Grassland-livestock balance constraints can include natural grassland utilization constraints, artificial grassland natural constraints, and non-negative constraints.

[0147] Natural grassland utilization constraints are constraints on the carrying capacity of natural grasslands, primarily requiring that livestock forage needs be matched with the forage yield of grassland resources in each season. Livestock grazing on natural grasslands should be less than the edible forage yield of natural grasslands, and the utilized area of ​​all types of natural grasslands should be less than the usable area of ​​natural grasslands.

[0148]

[0149]

[0150] in, It refers to the annual number of livestock in the pen. It refers to the number of various wild animals. The daily feed intake (kg / d) of various livestock and wild animals in the k-th type of natural grassland; The number of grazing days in natural grassland of type k (k=1 represents the warm season, which is 153 days; k=2 represents the cold season, which is 212 days). The usable grassland area is defined as type k natural grassland (k=1 represents warm-season pasture, k=2 represents cool-season pasture). ); The aboveground biomass of type k natural grassland ( ), This represents the area of ​​each planted artificial grassland. The yield per unit area of ​​each artificial grassland ( A represents the area of ​​the outer support zone.

[0151] Meanwhile, in order to achieve the goal of balanced nutrition for livestock, it is necessary to ensure that the supply of key nutrients such as crude protein and acid detergent fiber is precisely matched to the needs of livestock and wild animals, so that the intake of various nutrients can meet the needs of growth while avoiding metabolic imbalance caused by excessive supply, thereby achieving the best dynamic balance between nutrient supply and body demand.

[0152]

[0153]

[0154] In the formula: The daily crude protein requirement (kg / day) for various livestock and wild animals in the k-th type of natural grassland. Crude protein content (kg) of forage available from Weiran Ranch; The daily requirement (kg / day) of acid detergent fiber for various livestock and wild animals in the k-th type of natural grassland. The amount of acid detergent fiber (kg) that can be provided for natural pastures.

[0155] The natural constraints of artificial grasslands are determined by their carrying capacity. Ideally, the amount of supplemental feeding should be determined based on the actual conditions of the cool and warm seasons. Generally, supplemental feeding is not required in the warm season, but it is required in the cool season. During the cool season, supplemental feeding for livestock mainly consists of purchased hay, with the quantity purchased based on the number of livestock. If grazing is carried out during the cool season, the sum of the amount of hay planted on the artificial grassland and the amount of hay purchased should exceed the amount of supplemental feeding required for livestock.

[0156]

[0157] In the formula: h is the quantity of hay purchased; The area of ​​artificially planted grassland ( ); Hay yield of artificially planted grassland ( ), The amount of hay supplemented to livestock per day (kg / d). This refers to the number of days for supplemental feeding during the cold season.

[0158] Based on the "Calculation of Reasonable Carrying Capacity of Natural Grassland" (NY / T 635-2015) and related research, the theoretical carrying capacity (AGB) of grassland is calculated using the following formula ( ) and CP theory can provide quantity ( ):

[0159]

[0160] in, The crude protein content is the aboveground biomass (AGB), which is calculated from the nitrogen content of the plant sample. This represents the nitrogen content of the plant sample.

[0161]

[0162] U represents grassland utilization rate. In this embodiment of the application, the grassland utilization rate of the Qiangtang Nature Reserve can be set to 70%. This represents the theoretical supply during the warm or cool season, considering only the supply of grassland aboveground biomass (AGB).

[0163]

[0164] This indicates that the AGB (Active Grain Protein Standard) for warm or cool season grasslands theoretically provides the crude protein required. This refers to the number of days for grazing during the warm season, which is the growing season of natural grasslands. This refers to the number of days spent grazing during the cold season, not the growing season.

[0165] The amount of supplemental feed can be determined based on the carrying capacity, the aboveground biomass of livestock, and their crude protein intake requirements from plants.

[0166] Referring to the material and energy requirements of livestock in relevant studies, this application defines the grassland growing season (warm season) as the livestock weight gain season and the grassland withering season (cold season) as the physiological maintenance season with zero weight gain for livestock, thereby deriving different standards for the livestock requirements in the two seasons.

[0167] For example, since the grasslands in the Qiangtang Nature Reserve generally turn green in early May and wither in early October, this study defines the growing season in Qinghai Province as May to September, or 153 days, and the withering season as October to April of the following year, or 212 days. The actual carrying capacity of ungulate herbivorous wild animals was calculated by converting their equivalent body weight into corresponding standard sheep units: wild yak = 6 sheep units, Tibetan wild ass = 4 sheep units, white-lipped deer = 3 sheep units, Tibetan gazelle = 1.2 sheep units, and blue sheep = 1.1 sheep units.

[0168] The actual AGB demand for livestock and wildlife is obtained in the following manner ( ) and actual demand for CP ( ):

[0169]

[0170]

[0171] in, The actual daily demand of AGB for livestock and wildlife during the warm or cold seasons. This refers to the annual number of livestock and wild animals in stock. AGB daily food intake, representing livestock and wildlife AGB daily eclipse data representing livestock and wildlife during the warm season. AGB daily food intake represents the amount of food consumed by livestock and wildlife during the cold season; The actual daily demand for livestock and wildlife during the warm or cold seasons. This refers to the daily crude protein intake of livestock and wild animals. This refers to the daily crude protein intake of livestock and wild animals during the warm season. This refers to the daily crude protein intake of livestock and wild animals during the cold season. This refers to the number of days in the warm season, which is the growing season for natural grasslands. These are the days in the cool season, not the growing season.

[0172] When plants respond to biotic or abiotic stresses, they recover growth through resource redistribution mechanisms; this phenomenon is known as compensatory growth. Studies on the Qinghai-Tibet Plateau ecosystem have shown that grasslands in alpine steppe regions exhibit isocompensatory growth. Based on this, this study uses the peak biomass in the warm season as the upper limit for livestock forage supply, while the available resources in the cold season are half of the peak biomass in the warm season.

[0173] Based on two indicators—the availability of grassland and the actual demand of livestock and wildlife—the spatial distribution of areas with grass-livestock imbalance in the Qiangtang Nature Reserve and the amount of supplementary feed required to achieve a balance between forage supply and demand can be calculated.

[0174]

[0175]

[0176] in, The AGB forage amount is used to address the imbalance between grass and livestock during the warm or cold seasons in grasslands. This refers to the warm or cold season for grassland livestock. Combining the imbalanced CP (cumulative feed) with the above formula, the spatial distribution of the unbalanced grassland-livestock area can be determined from the raster image. Based on the Qiangtang Nature Reserve's research zoning (core protection zone, general control zone, and peripheral support zone) Shapefile file, the "Extract by mask" function in ArcGIS 10.7 can be used to obtain the required supplemental feed amounts in terms of aboveground biomass and crude protein nutrient indicators to achieve grassland-livestock balance in different areas.

[0177] The nonnegativity constraint requires that all variables in the model be nonnegative.

[0178] Please see Figure 4This is a spatial distribution map of aboveground biomass (AGB) and crude protein (CP) in the Qiangtang Nature Reserve, predicted in one embodiment of this application. Figure 4 As shown, the predicted results indicate that the aboveground biomass yield in the Qiangtang Nature Reserve ranges from 647.33 to 3186.96 kg / ha, with an average of 1107.02 ± 223.56 kg / ha. The total aboveground biomass yield is... The yield was lower in the northern part of the reserve and higher in the southern part, likely due to the higher altitude in the north. Crude protein yield ranged from 33.08 to 158.16 kg / ha, with an average of 84.26 ± 18.29 kg / ha. The total crude protein yield was... From northwest to southeast, crude protein production shows a clear increasing trend, with a distinct dividing line.

[0179] Please see Figure 5 This is a spatiotemporal distribution map of the seasonal imbalance between aboveground biomass supply and livestock in the Qiangtang Nature Reserve, as shown in one embodiment. Figure 5 As shown, during the warm season, the imbalance between grass and livestock biomass in the study area of ​​the Qiangtang Nature Reserve is relatively mild, spatially concentrated mainly in the core protected area, with fewer areas in the general control area and the outer support area. Areas with grass-livestock imbalance are mainly concentrated in the eastern part of the reserve and the southwestern part of the outer support area. During the cold season, there are more areas with grass-livestock imbalance in the reserve. The core protected area and the general control area are mostly experiencing mild imbalance, with a very few areas in the east requiring a larger amount of grass. In the outer support area, the amount of grass required gradually increases from north to south.

[0180] Please see Figure 6 This is a spatiotemporal distribution map of the seasonal imbalance between crude protein supply to livestock in the Qiangtang Nature Reserve, as shown in one embodiment. Figure 6 As shown, during the warm season, the imbalance of crude protein nutrient levels between grass and livestock in the study area of ​​the Qiangtang Nature Reserve is relatively mild, spatially concentrated mainly in the core protected area, with fewer areas in the general control area and the outer support area. Areas with crude protein nutrient imbalance between grass and livestock are mainly concentrated in the eastern part of the protected area and the western part of the outer support area. During the cold season, areas with crude protein nutrient imbalance between grass and livestock are mainly in the western part of the outer support area, with very few areas in the core protected area and the general control area, and some appearing in the eastern part of the region.

[0181] As shown in Table 8, the nutritional indicators required to achieve grass-livestock balance in the Qiangtang Nature Reserve vary in different seasons and regions. Combined with the spatiotemporal distribution, it can be seen that the nutritional requirements for grass-livestock balance exhibit significant spatiotemporal differentiation. The total grass requirement in the warm season is 7,497,623.608 kg, only 13.2% of the total requirement of 56,956,462.214 kg in the cold season, reflecting the unique seasonal division of the alpine grasslands: a long cold season and a short warm season. In the warm season, the core protected area requires the largest total grass supply, at 5,495,682.679 kg, while the outer support area requires the smallest, at 758,037.344 kg. Conversely, in the cold season, the outer support area requires the largest total grass supply, at 39,074,773.140 kg. From a regional structure perspective, the outer support area bears 68.6% of the forage supply pressure in the cold season, far exceeding its 10.1% share in the warm season.

[0182] Regarding the total crude protein requirement, the demand in the warm season is 349,498.971 kg, while the total demand in the cold season is 803,392.575 kg. The cold season demand is more than double that of the warm season, indicating the ecological characteristic of a sharp decline in the protein content of alpine grassland plants with the seasons. In the warm season, the core protected area requires the most total crude protein, while the outer support area requires the least total forage. The opposite is true in the cold season, with the trend consistent with the required forage supply. In the outer support area, the total crude protein in the warm season is 32,366.771 kg, accounting for 76.6% of the total crude protein demand in the warm season. In the cold season, the proportion of crude protein rises to 92.9%, indicating that this area plays a key role in maintaining the nutritional balance of livestock during winter. Relying solely on natural grasslands is insufficient to meet the nutritional needs of livestock during winter. It is necessary to refer to the experience of grass-livestock balance in the Yunnan-Guizhou region to develop a flexible supply model that coordinates artificial forage bases with natural grasslands.

[0183] Table 8. Number of nutritional indicators required for achieving grass-livestock balance in Qiangtang Nature Reserve

[0184]

[0185] Based on the above analysis, this application provides the following three optimization schemes:

[0186] As shown in Table 9, Option 1 strictly maintains the stability of wild animal populations (the increase in Tibetan wild ass, Tibetan antelope, and wild yak is controlled within a certain range). Under the premise of maintaining a growth rate within 2%, significant progress was made in improving habitat quality by 6.78% through optimizing livestock structure and forage resource allocation. Wild animal populations generally increased, with Tibetan wild asses increasing by 1.26%, Tibetan antelopes by 1.13%, and wild yaks by 0.80%, reflecting a systematic improvement in habitat quality. This is directly related to the 6.78% increase in total aboveground biomass (AGB). kg increased to kg, far exceeding the preset 5% threshold. At the same time, the unit of livestock product remained basically stable, decreasing by 0.46%; the income of pastoral areas decreased slightly by 1.31%, from 6,358,536,000 yuan to 6,275,455,988 yuan, while AGB feed intake increased slightly by 0.94%, possibly due to a slight increase in the number of wild animals.

[0187] While ensuring the annual forage demand is met, priority is given to meeting the demand during the cool season, as the imbalance between forage and livestock is more pronounced during this period. Supply during the warm season is then coordinated. During the cool season, the supply of AGB (Australian Grain Barrel) in the core protected area will be increased from its original value. kilograms increased to the planned value kilograms, an increase of 135.75%; CP supply from kilograms jumped to Kilograms (increase of 2895.48%, rate of change of 2995.48%). This indicates that the model improves by increasing artificial grassland planting (e.g., silage corn area increased by 306.40%) and hay purchases (from... kilograms increased to (kg, an increase of 65.91%), prioritizing the safety of livestock forage in the core protected area during winter to provide more survival resources for wildlife. The AGB supply in the general control area increased from... kilograms increased to kilograms, an increase of 10.55%; CP supply from kilograms increased to Kilograms (increase of 117.45%, change rate of 217.45%). The supply of AGB in the peripheral support zone from... kilograms increased to kilograms, an increase of 6.24%; CP supply from kilograms increased to kilogram.

[0188] With demand secured during the cold season, the model shifts to optimizing supply during the warm season to balance resource utilization and ecological restoration. During the warm season, the AGB supply in the core protected area increases from... kilograms increased to kilograms, an increase of 15.14%; CP supply from kilograms increased to kilograms, an increase of 11.18%. The supply of AGB in the general control area increased from... kilograms increased to kilograms, CP supply from kilograms increased to The model, by adjusting artificial grassland planting patterns, such as increasing triticale acreage by 496.20%, supplemented the warm-season forage gap. The AGB supply in the outer support zone increased from... kilograms increased to kilograms, an increase of 221.70%; CP supply from kilograms increased to Kilograms, an increase of 223.18%. The high growth in the warm season stems from resources released by the cool season-priority strategy, such as the reuse of hay reserves, which significantly improved the multifunctionality of grasslands. The total AGB (Amount in GB) increased from [amount in kg] to [amount in kg]. kilograms increased to kilograms, an increase of 6.78%.

[0189] Table 9. Multi-objective optimization results of Scheme 1

[0190]

[0191]

[0192] Option 1, while strictly maintaining stable wildlife populations, improved habitat quality by 6.78% through optimizing livestock structure and forage resource allocation, far exceeding the preset 5% threshold. Although pastoral income decreased slightly by 1.31%, total aboveground biomass (AGB) increased by 6.78%, reaching 88.58 × 10^8 kg, significantly improving habitat quality.

[0193] As shown in Table 10, Option 2 strictly maintains the stability of wild animal populations (the increase in Tibetan wild ass, Tibetan antelope, and wild yak is controlled within a certain range). Under the premise of (within 2%), the unit value of livestock products from the original value Increase to planned value The increase was 5.56%, while the total aboveground biomass (AGB), a core indicator of habitat quality, increased from [previous figure]. kilograms jumped to The weight gain was 10.89%, far exceeding the preset threshold.

[0194] Under the constraint of annual forage demand, the total AGB feed intake is from kilograms decreased slightly Kilograms, a decrease of 0.56%, prioritizing the alleviation of the grass-livestock imbalance during the cold season (historically severe nutritional gap), followed by optimization of the warm season supply. During the cold season optimization, the CP supply in the core protected area experienced an explosive increase of 2754.06% (from...). kilograms increased to (kg), AGB supply increased by 128.85% ( kilograms increased to (kg), thanks to a 33.35% increase in hay purchases (reaching) (kg) and livestock structure adjustment (such as an increase of 4.05% in the number of yaks slaughtered to (Head), which directly supports the survival resources for the growth of wildlife populations, such as the Tibetan wild ass population increasing by 1.99% to head.

[0195] During the warmer season, supply mainly fills the remaining gap, with AGB supply in the peripheral support zone surging by 222.64%. kilograms increased to (kg), CP supply increased by 113.15% (kg) kilograms increased to (kg), the general control area AGB also increased by 115.61% ( kilograms increased to (kg), the key driver is the precise allocation of artificial grassland planting (silage corn area +886.90%, oats +833.30%). Structural adjustments in livestock have brought controllable costs: sheep inventory decreased by 1.80% ( Head down The number of head (and yak) and the number of cattle decreased by 1.14%. Head down The reduction in grassland pressure (of the head) resulted in a slight decrease in pastoral income of 0.33%. Units reduced to (units), but this was completely offset by the growth in livestock products.

[0196] Table 10 Multi-objective optimization results of Scheme 2

[0197]

[0198]

[0199] Option 2, while maintaining stable wildlife populations, increased livestock product output by 5.56% and the total aboveground biomass (AGB), a core indicator of habitat quality, by 10.89%, achieving [the desired results]. kg. The imbalance between grass and livestock during the cold season has been effectively alleviated, and the income of pastoral areas decreased slightly by 0.33%, but this was completely offset by the increase in livestock products.

[0200] As shown in Table 11, Option 3 strictly maintains the stability of wild animal populations (the increase in Tibetan wild ass, Tibetan antelope, and wild yak is controlled within a certain range). Under the premise of (within 2%), the unit value of livestock products from the original value Increase to (Increase of 13.06%, change rate of 113.06%), the total aboveground biomass (AGB), a core indicator of habitat quality, increased from [previous figure]. kilograms jumped to The weight gain (in kilograms, an increase of 12.90%, a change rate of 112.90%) significantly exceeded the preset threshold. This was achieved while meeting the annual forage requirement (total AGB feed intake from...). kilograms dropped Under the constraint of (kg, a decrease of 1.23%), priority should be given to alleviating the imbalance between grass and livestock in the cold season, and then optimizing the supply in the warm season, so as to enhance ecological restoration while improving economic output.

[0201] During the cold season, the supply of CP in the core protected area exploded by 2713.83% (from kilograms increased to (kg), AGB supply increased by 119.85% ( kilograms increased to (kg), which is directly attributed to a 66.57% increase in hay purchases (reaching) (kg) and strategic adjustments to artificial grasslands (such as a surge of 898.00% in silage corn acreage). Such optimizations significantly reduced the wintering risks for livestock and wildlife (Tibetan antelope population increased slightly by 0.26% to (head), although the Tibetan wild ass population decreased slightly by 1.71% (to) (Head), but the overall improvement in habitat quality indicates enhanced system resilience. CP supply in the general control area increased by 113.14% (head), but the overall improvement in habitat quality indicates enhanced system resilience. kilograms increased to (kg), AGB supply increased by 17.38% (kg) kilograms increased to (kg), through adjustments to livestock structure (such as an increase of 13.17% in the number of yaks slaughtered to...) (Head) Relieves pressure on grasslands and ensures a dynamic balance between nutrient supply and demand. The CP supply in the outer support zone saw a steady increase of 12.55%. kilograms increased to (kg), AGB supply increased by 3.44% ( kilograms increased to (kg), providing a buffer for basic ecological security during the cold season, supporting a 0.85% increase in the wild yak population (to... head).

[0202] After ensuring demand during the cold season, the model shifts to resource reallocation during the warm season to fill the remaining gap and amplify ecological benefits. During the warm season, the supply of AGB in general control areas surges by 84.74% ( kilograms increased to (kg), CP supply increased by 7.69% (kg). kilograms increased to (kg), the key driver was the precise allocation of artificial grasslands (triticale area increased by 894.60%, oat area increased by 898.30%), effectively supplementing the forage reserves optimized for the cool season. AGB supply in the peripheral support areas increased by 77.42% ( kilograms increased to (kg), CP supply surged 189.42% ( kilograms increased to (kg), thanks to a 799.40% expansion in the area of ​​*Vicia sativa*, unlocking the multifunctional potential of warm-season grasslands. AGB supply in the core protected area increased moderately by 1.65% (kg). kilograms increased to (kg), CP supply increased by 4.21% (kg) kilograms increased to (kilograms), reflecting the resource spillover effect and ensuring the continuity of wildlife habitats (such as the stable growth of wild yak populations).

[0203] To achieve the dual-objective leap, the model adjusted the livestock structure, moderately reducing the number of livestock, with the number of sheep decreasing by 1.24%. Head down (head), the number of yaks decreased by 1.87% ( Head down The increased grassland carrying capacity (a slight increase of 0.05% in the usable area of ​​natural grassland) directly enhances ecosystem services. Pastoral income increased by 11.67%. Unit increased to The unit (unit) completely offset the 1.23% decrease in AGB feed intake, proving the feasibility of "promoting the economy through ecology".

[0204] Table 11 Multi-objective optimization results of Scheme 3

[0205]

[0206]

[0207] Option 3, while strictly maintaining the stability of wild animal populations, achieved a 13.06% increase in livestock product per unit, a 12.90% increase in total aboveground biomass (AGB), a core indicator of habitat quality, and an 11.67% increase in pastoral income, completely offsetting the decline in AGB feed intake.

[0208] Based on the above research results, the following planning recommendations can be made: (1) Strengthen the construction of artificial grasslands. The government should increase its support for the construction of artificial grasslands and promote the planting of high-yield, high-protein forage crops, especially in areas where there is insufficient forage supply in the cold season, to ensure the nutritional needs of livestock during winter. (2) Optimize the herd structure. Herdsmen should be encouraged to adjust their herd structure, appropriately increase the proportion of female livestock, and improve the efficiency of slaughtering, so as to reduce the grazing pressure on grasslands and increase income from livestock products. (3) Protect wildlife habitats. Livestock grazing areas should be rationally planned within protected areas to avoid excessive overlap with wildlife habitats and ensure that wildlife have sufficient living space and food sources. (4) Promote the ecological compensation mechanism. A sound ecological compensation mechanism should be established to encourage herdsmen to participate in ecological protection and promote the coordinated development of the social economy and ecological environment in pastoral areas through policy incentives and economic compensation.

[0209] This application proposes a multi-objective optimization method for nature reserve resources based on the balance of "human settlement-grassland-wildlife-livestock". This method extends the traditional grassland-livestock balance to a broader ecosystem management framework, considering not only the coexistence of livestock and wildlife but also the livelihoods of herders and the sustainable development of grassland ecosystems. Through scientific planning and rational utilization of natural resources, it aims to increase herders' income while protecting the health and diversity of grassland ecosystems, achieving a win-win situation for ecological protection and economic development. Through regional and seasonal model simulations, the introduction of CP nutrient indicators for balance, and an innovative multi-objective optimization model, this method provides new ideas and approaches for the management and protection of alpine grassland ecosystems in nature reserves.

[0210] This application focuses not only on aboveground biomass (AGB) in protected areas but also, and more specifically, on the balance of crude protein (CP), a key nutrient indicator. Protein is an essential nutrient for maintaining the health and reproductive capacity of livestock and wildlife. Traditional grassland-livestock balance studies in protected areas often limit themselves to the supply and demand of aboveground biomass (AGB), neglecting the needs of livestock and wildlife for key nutrients such as protein. In the alpine grassland ecosystem of the Qinghai-Tibet Plateau, the protein content of forage grasses varies dramatically with the seasons, with significantly lower protein content in cold-season forage grasses compared to warm-season grasses. This poses a significant challenge to the nutritional needs of livestock and wildlife. This study introduces the crude protein (CP) indicator to construct a nutrient-based grassland-livestock balance model, enabling a more comprehensive assessment of the carrying capacity of grassland ecosystems and the nutritional needs of livestock and wildlife. For example, lower CP content in cold-season forage grasses leads to insufficient nutrient intake for livestock and wildlife, thus affecting their growth and reproduction. By predicting the crude protein (CP) supply in different regions and seasons and combining it with the predicted aboveground biomass (AGB), this study ensures that livestock and wildlife can obtain sufficient nutrition even in the cold season. Furthermore, the introduction of crude protein (CP) provides a new perspective for evaluating the effectiveness of artificial grassland planting. For example, certain high-protein forage crops (such as silage corn, triticale, and oats) have played an important role in supplemental feeding during the cool season, significantly improving the nutritional intake levels of livestock and wild animals.

[0211] Traditional studies often focus on the overall assessment of the entire protected area, neglecting the differences between different regions and seasons. This can lead to uneven resource allocation and reduced effectiveness of management measures. This application employs a regional and seasonal approach to model the protected area, enabling more precise formulation of supplemental feeding strategies. Particularly in terms of seasons, the study meticulously divides the grassland growing season (warm season) and the withering season (cool season). The warm season (May to September) is crucial for grassland greening and livestock weight gain, while the cool season (October to April of the following year) is the period for livestock to maintain their physiological needs. Detailed simulation of these two seasons allows for a better understanding of the dynamic changes in the grassland ecosystem, particularly the grass-livestock imbalance during the cool season. For example, the grass-livestock imbalance is particularly pronounced during the cool season because the supply of natural pasture is insufficient, making it difficult to meet the nutritional needs of livestock and wildlife. Seasonal simulations allow for more accurate prediction of grass-livestock imbalance areas during the cool season, enabling proactive measures such as increasing artificial grassland planting and purchasing hay to ensure the safe overwintering of livestock and wildlife.

[0212] This application employs a deep neural network model to perform high-precision spatial prediction of grassland aboveground biomass (AGB) and crude protein (CP). Compared with traditional statistical models, deep neural network models can better handle complex nonlinear relationships, especially when dealing with high-dimensional and heterogeneous data. Through feature selection and model optimization using the Boruta algorithm, a high-precision AGB and CP prediction model was constructed, providing a solid data foundation for subsequent multi-objective optimization. The non-dominated sorting genetic algorithm (NSGA-II) is used to optimize multiple objectives such as herd structure, wildlife population, and artificial grassland planting. The NSGA-II algorithm can find the optimal solution set (i.e., Pareto front) among multiple objectives, thus avoiding the problems of premature convergence and low computational efficiency that may occur with traditional weighted sum methods. This allows for planning information that can improve herders' income while protecting the health and diversity of the grassland ecosystem.

[0213] like Figure 7 As shown in the illustration, this application also provides a multi-objective optimization device for resources in nature reserves, the device comprising:

[0214] The data acquisition module 110 is used to acquire biological data and multi-source remote sensing data of the target nature reserve; wherein, the biological data includes grassland data, livestock data and wildlife data;

[0215] The data prediction module 120 is used to obtain aboveground biomass data and plant crude protein data of the target nature reserve based on the multi-source remote sensing data and the pre-trained prediction model.

[0216] The decision information acquisition module 130 is used to acquire the decision objectives and constraints of the target nature reserve; wherein, there are at least two decision objectives, and the decision objectives and constraints are related to at least one of the aboveground biomass data, crude protein data, grassland data and animal data;

[0217] The planning information acquisition module 140 is used to acquire the optimal planning information of the target nature reserve based on the second-generation non-dominated sorting genetic algorithm, according to the biological data of the target nature reserve, the aboveground biomass data, the plant crude protein data, the decision objective, and the constraints; wherein, the optimal planning information includes livestock structure, wild animal numbers, and artificial grassland planting information.

[0218] It should be noted that the multi-objective optimization device for nature reserve resources provided in the above embodiments is only illustrated by the division of the above functional modules when executing the multi-objective optimization method for nature reserve resources. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the multi-objective optimization device for nature reserve resources provided in the above embodiments and the multi-objective optimization method for nature reserve resources in the above embodiments belong to the same concept, and its implementation process is detailed in the method embodiments, which will not be repeated here.

[0219] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the multi-objective optimization method for resources in nature reserves as described in any of the above embodiments.

[0220] The embodiments of this application may take the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing program code. Computer-readable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information may be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to: phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0221] like Figure 8 As shown, this application embodiment also provides a computer device 200, including a memory 210, a processor 220, and a computer program stored in the memory 210 and executable by the processor 220;

[0222] When the processor 220 executes the computer program, it implements the steps of the multi-objective optimization method for nature reserve resources as described in any of the above.

[0223] The memory 210 includes read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0224] The processor 220 is the control unit of the computer device 200. It connects to various components of the computer device 200 via various interfaces and lines. By running or executing programs or modules stored in the memory 210, and by calling data stored in the memory 210, it performs various functions of the computer device 200 and processes data. For example, when the processor 220 executes the computer program stored in the memory 210, it implements all or part of the steps of the multi-objective optimization method for nature reserve resources described in this application embodiment; or it implements all or part of the functions of the multi-objective optimization device for nature reserve resources. The processor 220 can be composed of integrated circuits, such as a single packaged integrated circuit, or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips.

[0225] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0226] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A multi-objective optimization method for resources in nature reserves, characterized in that, include: Acquire biological data and multi-source remote sensing data of the target nature reserve; wherein the biological data includes grassland data, livestock data and wildlife data; Based on the multi-source remote sensing data and the pre-trained prediction model, obtain aboveground biomass data and plant crude protein data of the target nature reserve; The decision-making objectives and constraints of the target nature reserve are obtained; wherein, the decision-making objectives include grassland ecological environment optimization objectives; and the constraints include livestock production constraints and grassland-livestock balance constraints. The objectives for optimizing the grassland ecological environment include minimizing grassland resource pressure: )， in, Pressure on grassland resources This represents the year-end inventory of each type of livestock. For each livestock, the amount of above-ground organisms consumed; For the number of each wild animal, This refers to the daily food intake of various wild animals. Let k be the number of grazing days for natural grassland of type k. When k=1, the corresponding number of grazing days for natural grassland in the warm season is 153 days, and when k=2, the corresponding number of grazing days for natural grassland in the cool season is 212 days. The livestock production constraints are used to keep the slaughter rate of the corresponding livestock herd within a set range: ,in, and These are the upper and lower limits of the slaughter rate; The grassland-livestock balance constraints include natural grassland utilization constraints and artificial grassland natural constraints. The natural grassland utilization constraints are used to match livestock forage requirements with the forage yield of grassland resources in each season: ， ， in, It refers to the annual number of livestock in the pen. It refers to the number of various wild animals. The daily feed intake of various livestock and wild animals in the k-th type of natural grassland; Let k be the number of grazing days for natural grassland of type k. When k=1, the corresponding number of grazing days for natural grassland in the warm season is 153 days, and when k=2, the corresponding number of grazing days for natural grassland in the cool season is 212 days. Let k be the usable grassland area of ​​natural grassland of type k, where k=1 is warm-season pasture and k=2 is cool-season pasture. This represents the aboveground biomass of type k natural grassland. This represents the area of ​​each planted artificial grassland. For the yield per unit area of ​​each artificial grassland, The area of ​​the outer support zone; The natural constraints of the artificial grassland are the constraints of the artificially planted grassland: ， in, The quantity of hay purchased; This refers to the area of ​​artificially planted grassland; The dry hay yield of artificially planted grasslands. The amount of hay to supplement livestock's daily feed. Number of days for supplemental feeding during the cold season; Based on the second-generation non-dominated sorting genetic algorithm, a nutrient-based grassland-livestock balance model is constructed according to the biological data of the target nature reserve, the aboveground biomass data, the plant crude protein data, the decision objective, and the constraints, to obtain the optimal planning information for the target nature reserve; wherein, the optimal planning information includes herd structure, wildlife population, and artificial grassland planting information; including: Establishing the optimal dynamic balance between nutrient supply and the body's needs: ， ， in, To determine the daily crude protein requirements of various livestock and wild animals in the k-th type of natural grassland, The amount of crude protein that natural pastures can provide; The daily acid detergent fiber requirements for various livestock and wild animals in the k-th type of natural grassland. The amount of acid detergent fiber that natural pastures can provide; Based on the availability of grassland and the actual demand of livestock and wildlife, the spatial distribution of areas with grassland-livestock imbalance and the amount of supplemental feed required to achieve forage supply-demand balance were calculated: ， ， in, This refers to the aboveground biomass caused by the imbalance between grass and livestock during the warm or cold seasons of grassland. The actual daily biomass requirements of livestock and wildlife during the warm or cold seasons. This provides theoretical estimates of the amount of aboveground biomass that can be generated in warm or cool season grasslands. The crude protein content is related to the imbalance between grass and livestock during the warm or cool seasons of grassland. This refers to the actual daily crude protein requirements of livestock and wild animals during the warm or cold seasons. The theoretical amount of crude protein that can be provided for grassland biomass during the warm or cool seasons.

2. The multi-objective optimization method for resources in nature reserves according to claim 1, characterized in that, The multi-source remote sensing data includes at least two variables, and after acquiring the multi-source remote sensing data, it includes: Based on the Pearson correlation analysis algorithm, relevant variables that are correlated with aboveground biomass and plant crude protein are selected from the at least two variables; Based on the Boruta feature selection algorithm, significant variables that are significantly correlated with the aboveground biomass and the plant crude protein are selected from the relevant variables to obtain a feature subset; Sampling points were randomly selected in the target nature reserve to obtain aboveground biomass and plant crude protein data, thus obtaining sample data. A dataset is generated based on the feature subset and sample data. The dataset is then used to pre-train a deep neural network model to obtain a prediction model.

3. The multi-objective optimization method for resources in nature reserves according to claim 1, characterized in that, Based on the multi-source remote sensing data and the pre-trained prediction model, aboveground biomass data and plant crude protein data of the target nature reserve are obtained, including: The multi-source remote sensing data is preprocessed; wherein, the preprocessing includes missing value handling, feature standardization, and spatial alignment; The spatially heterogeneous raster data in the multi-source remote sensing data are uniformly resampled and coordinate transformed to unify the raster data to the target resolution; Several single-band raster data are read block by block, and the pre-trained prediction model is used to generate the aboveground biomass prediction value and plant crude protein prediction value per pixel.

4. The multi-objective optimization method for resources in nature reserves according to claim 1, characterized in that, The decision-making objectives include livestock productivity optimization objectives, which include increasing the slaughter rate: in, For slaughter rate, This represents the annual sales volume of livestock i. Let j be the number of livestock at the end of the year.

5. The multi-objective optimization method for resources in nature reserves according to claim 1, characterized in that, The decision-making objective includes an economic benefit optimization objective, which includes maximizing livestock output value. in, For livestock output value, This represents the average annual output value of meat, wool, and cashmere per sheep. The annual output value per goat, The annual output value per yak For the number of sheep, For the number of goats, For the number of yaks, This represents the average price of purchasing 1 kg of hay. To determine the quantity of hay to purchase, This represents the area of ​​each planted artificial grassland. For planting 1 hm 2 The cost.

6. The multi-objective optimization method for resources in nature reserves according to claim 1, characterized in that, The decision-making objectives include grassland habitat quality objectives, which are used to optimize the aboveground biomass of natural and artificial grasslands to achieve ecosystem balance and stability. in, The habitat quality is represented by u, the total usable area of ​​natural grassland, and d, the mean aboveground biomass of the protected area. This represents the area of ​​each planted artificial grassland. This represents the yield per acre of each artificial grassland.

7. A multi-objective optimization device for resources in nature reserves, characterized in that, include: The data acquisition module is used to acquire biological data and multi-source remote sensing data of the target nature reserve; wherein, the biological data includes grassland data, livestock data and wildlife data; The data prediction module is used to obtain aboveground biomass data and plant crude protein data of the target nature reserve based on the multi-source remote sensing data and the pre-trained prediction model. The decision information acquisition module is used to acquire the decision objectives and constraints of the target nature reserve; wherein, the decision objectives... This includes objectives for optimizing the grassland ecological environment; the constraints include constraints on livestock production and constraints on the balance between grassland and livestock. The objectives for optimizing the grassland ecological environment include minimizing grassland resource pressure: )， in, Pressure on grassland resources This represents the year-end inventory of each type of livestock. For each livestock, the amount of above-ground organisms consumed; For the number of each wild animal, This refers to the daily food intake of various wild animals. Let k be the number of grazing days for natural grassland of type k. When k=1, the corresponding number of grazing days for natural grassland in the warm season is 153 days, and when k=2, the corresponding number of grazing days for natural grassland in the cool season is 212 days. The livestock production constraints are used to keep the slaughter rate of the corresponding livestock herd within a set range: ,in, and These are the upper and lower limits of the slaughter rate; The grassland-livestock balance constraints include natural grassland utilization constraints and artificial grassland natural constraints. The natural grassland utilization constraints are used to match livestock forage requirements with the forage yield of grassland resources in each season: ， ， in, It refers to the annual number of livestock in the pen. It refers to the number of various wild animals. The daily feed intake of various livestock and wild animals in the k-th type of natural grassland; Let k be the number of grazing days for natural grassland of type k. When k=1, the corresponding number of grazing days for natural grassland in the warm season is 153 days, and when k=2, the corresponding number of grazing days for natural grassland in the cool season is 212 days. Let k be the usable grassland area of ​​natural grassland of type k, where k=1 is warm-season pasture and k=2 is cool-season pasture. This represents the aboveground biomass of type k natural grassland. This represents the area of ​​each planted artificial grassland. For the yield per unit area of ​​each artificial grassland, The area of ​​the outer support zone; The natural constraints of the artificial grassland are the constraints of the artificially planted grassland: ， in, The quantity of hay purchased; This refers to the area of ​​artificially planted grassland; The dry hay yield of artificially planted grasslands. The amount of hay to supplement livestock's daily feed. Number of days for supplemental feeding during the cold season; The planning information acquisition module is used to construct a nutrition-based grassland-livestock balance model based on the biological data, aboveground biomass data, plant crude protein data, decision objectives, and constraints of the target nature reserve using a second-generation non-dominated sorting genetic algorithm, thereby obtaining optimal planning information for the target nature reserve. The optimal planning information includes livestock herd structure, wildlife populations, and artificial grassland planting information. Establishing the optimal dynamic balance between nutrient supply and the body's needs: ， ， in, To determine the daily crude protein requirements of various livestock and wild animals in the k-th type of natural grassland, The amount of crude protein that natural pastures can provide; The daily acid detergent fiber requirements for various livestock and wild animals in the k-th type of natural grassland. The amount of acid detergent fiber that natural pastures can provide; Based on the availability of grassland and the actual demand of livestock and wildlife, the spatial distribution of areas with grassland-livestock imbalance and the amount of supplemental feed required to achieve forage supply-demand balance were calculated: ， ， in, This refers to the aboveground biomass caused by the imbalance between grass and livestock during the warm or cold seasons of grassland. The actual daily biomass requirements of livestock and wildlife during the warm or cold seasons. This provides theoretical estimates of the amount of aboveground biomass that can be generated in warm or cool season grasslands. The crude protein content is related to the imbalance between grass and livestock during the warm or cool seasons of grassland. This refers to the actual daily crude protein requirements of livestock and wild animals during the warm or cold seasons. The theoretical amount of crude protein that can be provided for grassland biomass during the warm or cool seasons.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the multi-objective optimization method for resources in nature reserves as described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable by the processor; When the processor executes the computer program, it implements the steps of the multi-objective optimization method for resources in nature reserves as described in any one of claims 1-6.

Images