Grassland resource evaluation based on improved bp network model and analytic hierarchy process

Abstract

Grassland resources are an important part of land resources. Moreover, it has the functions of regulating the climate, windproof and sand fixation, conserving water sources, maintaining water and soil, raising livestock, providing food, purifying the air, and beautifying the environment in terrestrial ecosystems. Grassland resource evaluation is of great significance to the sustainable development of grassland resources. Therefore, this paper improves the BP neural network, uses the comprehensive index method to calculate the weights in the analytic hierarchy process, and constructs a water resources carrying capacity research and analysis system based on the entropy weight extension decision theory. Meanwhile, this paper analyzes different levels of resource and environmental carrying capacity to achieve the purpose of comprehensive evaluation of resource and environmental carrying capacity. In addition, based on the theory of sustainable development, under the guidance of the principle of index system construction, this paper studies the actual situation of grassland resources and the availability and operability of data, and combines with the opinions given by experts to form an evaluation index system of grassland resources and environmental carrying capacity. Finally, through the actual case study analysis, it is concluded that the model constructed in this paper has a certain effect.

Keywords

BP neural network improved algorithm analytic hierarchy process grassland resources

1 Introduction

China is one of the countries with the richest grassland resources in the world, and the total area of natural grassland in our country is nearly 400 million hm2. Moreover, grassland is the largest biological resource and green vegetation covering my country’s land area, accounting for about 42% of my country’s land area, which is equivalent to 4 times the cultivated land area and 3 times the forest land area. Therefore, it has important ecological, economic and social cultural values. Grassland resources are the material basis for the sound development of animal husbandry and an important part of my country’s economic development. In addition, the development of animal husbandry production must first develop grassland resources, and the development of grassland resources is based on grassland. Grassland resources occupy a considerable proportion in my country, mainly concentrated in pure pastoral areas or semi-agricultural and semi-pastoral areas in northwestern China. Meanwhile, grassland resources provide food for domestic animals and wild animals and provide a good living environment and habitat for wild animals and plants. At the same time, grassland resources are the material basis on which the people in pastoral areas depend, and the base for local people to produce and live. Therefore, grassland resources play an important role in animal husbandry production and national economy. In addition, grassland resources are an important part of land resources, and it has the functions of regulating climate, windproof and sand fixation, conserving water sources, maintaining soil and water, raising livestock, providing food, purifying air, and beautifying the environment in terrestrial ecosystems [1].

At present, from the perspective of land resource research topics at home and abroad, it can be seen that there are more researches on agricultural land and urban land, but less research on pasture. In the research of grassland resources, there are relatively more researches on the changes of grassland property rights system, the relationship between property rights system and grassland degradation, and less researches on the sustainable use of grassland resources. So far, there have been few researches on the sustainable use of grassland in Chayouzhongqi, so it is of great significance to study the sustainable use of grassland resources in Chayouzhongqi.

In this paper, a comprehensive evaluation of the sustainable use of grassland resources in Chayouzhong Banner is made based on the current status of local land resource utilization and grassland resource utilization. Moreover, this paper obtains the sustainable utilization status of grassland resources through evaluation, which has certain theoretical significance for establishing and improving the research framework system of grassland resource management in the future [2].

Grassland resources are an important part of land resources, and grassland is also the base of animal husbandry production, the barrier of the ecological environment, and it has a great relationship with the socio-economic and environmental sustainable development, and its role is irreplaceable. Chayouzhongqi, located in the back mountain area of Ulanqab, is located in the middle and western parts of the country with severe environmental conditions. Under the environmental conditions of population pressure and inflated demand, due to the problems of long-term overgrazing, over-cultivation, over-cutting, and construction speeds far lower than the degradation rate, 90% of the grasslands in Chayouzhongqi area have been or are degrading. Therefore, to prevent grassland degradation and desertification, effective use of grassland resources is an effective way to achieve sustainable development of grassland resources [3].

2 Related work

Land use change not only shows changes in quantity and quality, but also changes in spatial pattern. Spatial pattern and its changes are important aspects of land use change research and a key issue in the field of land use change research. With the deepening of the research, the research on the status of LUCC has gradually shifted from a simple status quo study to an analysis of the spatial pattern. After years of development, the LUCC research plan has been widely developed and implemented in various countries around the world. The research content also extends from the research on the effects of global climate change to the land use and land cover change processes at different spatial scales, driving mechanisms, and the study of the effects of resources, ecology, and environmental effects [4]. The research content of LUCC around the world is rich and involves many fields. These studies select an appropriate case study area on a certain time and space scale and conduct a research on a certain LUCC scientific issue based on the theory and method used to obtain various basic data. Among them, the theory and method of research, the choice of case area, the spatial and temporal scale of the research, and the acquisition of basic data and information all belong to the foundation of LUCC research, and the scientific issues studied are the core content of LUCC research. The fundamental starting point and basic goals of the LUCC research plan are: by understanding the dynamics (dynamic processes) and interaction mechanisms between human driving forces, land use and land cover change, global change and environmental feedback, a LUCC model that can be used to predict land use and land cover change assessment and environmental change assessment and provide decision support is established, thereby predicting land use and land cover changes, assessing changes in the ecological environment, and seeking active human intervention measures. The literature [5] believed that the current hotspot in the field of LUCC research lies in the research and evaluation of the environmental effects of LUCC. With the development of the LUCC research mechanism, there has been considerable accumulation of basic data on LUCC research. Based on the research on the dynamic changes of land use and land cover, change mechanisms and driving factors, driving mechanisms, etc., the focus of LUCC research has gradually shifted to effects. Its main contents include the ecological effects and evaluation of land use and land cover changes, ecosystem response and health evaluation, and the ecological effects caused by a certain natural element (mainly soil, hydrology, climate, etc.) under land use and land cover changes. The literature [6] discussed the mechanism of the terrestrial ecosystem’s response to global changes and used the Holdridge Vegetation-Climate Classification System to simulate the response of Chinese vegetation to global changes, and to make predictions of changes. Moreover, it has made useful explorations of LUCC changes and feedback on global changes, especially climate change. The literature [7] discussed the change mechanism of LUCC and the mechanism of ecological environment through the study of land quality. Moreover, it studied how land use and land cover changes affect many natural factors such as soil, climate, hydrology, and biogeochemical cycles. In addition, it believed that studying the mechanism of action of LUCC is of great significance to better grasp and predict the ecological effects caused by LUCC.

The qualitative method and the quantitative method are the main methods of grassland resource evaluation. The qualitative method uses a combination of personal experience and knowledge to evaluate grassland resources, and it mainly includes experience judgment method and expert consultation method. The quantitative method is based on the qualitative method and relies on various factors and evaluation results to evaluate grassland resources. The commonly used methods are analytic hierarchy process, gray correlation method, index method, comprehensive evaluation of fuzzy mathematics, and multivariate analysis method [8].

At present, grassland resource evaluation methods gradually change from qualitative research to quantitative research and use computer-aided technology for evaluation. The grassland resource comprehensive evaluation unit can target grassland division units or grassland types.

The literature [8] tested the impact on grasslands under different grazing methods and came to the conclusion that the control of grazing intensity on pastures of different quality, grass yield, and yield is the only way to get the maximum output. The literature [9] studied the economic benefits of winter lamb production. The literature [10] conducted a comparative study on the livestock production capacity of different grazing conditions and summarized a set of best feeding and management methods. The literature [11] studied the supplementary feeding of winter lambs. The literature [12] believed that the combination of winter grazing and captivity is a necessary means to increase farmers’ income. The literature [13] put forward a plan for optimizing the herd and stock structure of grassland resources. The literature [14] studied the contradiction between livestock and grass in Ankang City and summarized the configuration of animal species structure and the best herd turnover pattern, which provided a reference for the development of local animal husbandry. The literature [15] studied the grazing management system of grassland and goats and concluded that the implementation of the optimization model of feeding management, the improvement of the herd turnover model, and the implementation of the forage quota system can obtain the maximum benefit.

The literature [16] proposed the “three belts, three seasons, one change” grassland optimal allocation model, which has become the guiding principle of Xinjiang grassland resources optimal allocation. The literature [17] used “3S” technology to study the grassland resource utilization pattern of Zhaosu military horse farm. The literature [18] believed that to sustainably develop grassland resources, it is necessary to increase the herders’ awareness of protecting the ecology, enhance their responsibility to protect the ecology, and protect biodiversity. The literature [19] established a model for estimating grassland degradation and stocking capacity, and through the study of grassland production performance and forage nutrition status, summed up the production model of sustainable animal husbandry. The literature [20] studied the production performance of grassland and livestock in Horqin grassland and used organic control technology to organically combine the production, supply and sales of grassland nuclear animal husbandry, and used linear programming to optimize the allocation of animal husbandry production structure. The article [26] implementated IoT-based Smart City is achieved by exploiting IoT and BigData Analytics using Hadoop ecosystem in real time environments. The article [27] reflects on IoT and its main role in the development of human behaviors and actions. The paper also deals with the compilation of various data from different databases connected to the Internet. The literature [28] addresses the numerous issues in the field of vehicle communication with the suggestion for a mutual unified and dispersed spectrum sensing model. The introduction of a mutual cognitive paradigm minimizes conflict and multiple unknown problems. The literature [29] discusses the issue, such as large amount of bigdata, and introduces the SmartBuddy framework for creating smart and adaptive ecosystems using human behaviors and human dynamics. The article [30] talks around the development of coordinated non-cyclic chart for video coding calculations for movement estimation in parallel reconfigurable computing frameworks [31]. The partitioning algorithm moreover plays a key part in optimizing the encoding of images [32].

3 Theoretical basis for the establishment of evaluation models

The standardization of indicators is the dimensionless process of evaluating indicators. The indicators selected in the evaluation are not uniform in nature, so they are not directly comparable and cannot be directly calculated in the comprehensive evaluation process. Therefore, the index data needs to be converted to a unified standard through a certain method to eliminate the dimension. The standardization process of each index in this study adopts the method of range standardization, and the standardization value of the index is controlled between [0, 1]. Generally speaking, indicators can be divided into three categories according to the evaluation target orientation: the indicator value “bigger is better”, “smaller is better” and “moderate is appropriate”, which corresponds to positive indicators, negative indicators and neutral index. Positive index means that the larger the index value, the greater its impact on the evaluation target. negative index means the smaller the index value, the greater its impact on the evaluation target. Meanwhile, neutral index refers to between the positive index and the negative index, the index value should not be large or small. This research indicator system contains only positive indicators and negative indicators, and the standardized formula is as follows:

Positive indicator means that the larger the indicator value, the better, and its calculation formula is as follows [21]: $p_{ij} = {\frac{x_{ij} - min_{j} x_{ij}}{max_{j} x_{ij} - min_{j} x_{ij}}}$ (1)

Negative indicator means that the smaller the indicator value, the better, and its calculation formula is as follows: $p_{ij} = {\frac{max_{j} x_{ij} - x_{ij}}{max_{j} x_{ij} - min_{j} x_{ij}}}$ (2)

In the formula, p_ij represents the value of the standard after eliminating the dimension, x_ij represents the calculation result of the original data of the target, and i represents the serial number of the indicator. At the same time, j represents the serial number of the evaluation area.

Index weight is a measure of the importance of each index in the index system relative to the target object, and different weights will bring different evaluation results. Therefore, scientific and reasonable allocation of index weights is very important for grassland resource distribution and utilization evaluation. The subjective weighting method and the objective weighting method are common methods for determining weights [22].

The subjective weighting method refers to people’s subjective determination of the weight of each factor of the analysis object according to its importance and experience. This type of method is more mature, but it is less objective and has higher requirements for evaluators, such as AHP method and expert scoring method. The objective weighting method refers to sorting, calculating and analyzing the actual data to obtain the weight, which avoids the subjectivity of the subjective weighting method to a certain extent. However, the calculation method of the objective weighting method is complicated, ignoring the different importance performance of different indicators relative to the target object. By considering the advantages and disadvantages of the subjective weighting method and the objective weighting method, this study uses a combination of subjective weighting method and objective weighting method to determine the index weight. The combined weighting method takes into account not only people’s experience and knowledge, but also objective information of the original data. Moreover, through the combination of subjective weighting method and objective weighting, the evaluation results are more accurate.

The analytic hierarchy process is a more commonly used method in the subjective weighting method. It has the advantages of requiring less data, short calculation time, and less workload, so it is suitable for grassland resource distribution and utilization evaluation and reflects the integrity of the evaluation objectives. The entropy weight method is one of the more reliable methods of evaluation results in the objective weighting method. It is an objective evaluation and analysis of a certain index system given the evaluation object. The obtained index weight is the response of each index based on the relative influence degree of a certain set of data in a certain sense. Therefore, the entropy weight method has been widely used in scheme optimization, multi-objective decision-making and various comprehensive evaluations. Moreover, its application field involves almost all disciplines such as engineering technology, management science, social economy and decision theory.

Due to the different degree of influence of each component on the overall state of grassland resources, in the comprehensive evaluation of the distribution and utilization of grassland resources, each component should have different weights. Therefore, this study selects the combined weight method to calculate the weight of grassland resource distribution and utilization. It is a weighting method that combines the analytic hierarchy process of subjective weighting method and the entropy weighting method of objective weighting method [23].

1. Analytic Hierarchy Process

First, this paper determines the scores of the relative importance of each level of indicators based on the opinions of experts, and then uses the Analytic Hierarchy Process (AHP) to determine the weight of each indicator. Specific steps are as follows:

The hierarchical model of the target layer structure is established.

The hierarchical structure designed in this paper are: target layer, criterion layer and index layer.

The judgment matrix is constructed.

A is the goal, u_i, u_j (i, j = 1, 2, ⋯ , n) is the factor, and u_ij is the relative importance of u_i to u_j. Moreover, this paper uses u_ij to form the A - P judgment matrix P. $P = [\begin{matrix} U_{11} & U_{12} & \dots & U_{1 n} \\ U_{21} & U_{22} & \dots & U_{2 n} \\ ⋮ & ⋮ & \dots & ⋮ \\ U_{n 1} & U_{n 2} & \dots & U_{nn} \end{matrix}]$ (3)

For the determination of the value of u_ij in the matrix, drawing on the method of Santy et al., the number l–9 and its reciprocal are used as the scale. After the analysis level is established, the subordination relationship between each level is clear, and a pairwise judgment matrix of different levels is established. The principle of the AHP method is that except for the highest layer, each layer must establish a judgment matrix, and the number of the matrix is equal to the element value of the previous layer. After that, this paper uses the 1–9 scale method to establish the judgment matrix u_ij, which meets the requirements of u_ij > 0 and u_ij = 1/u_ij.

The importance ranking is calculated.

According to the judgment matrix, the feature vector ω corresponding to the largest feature root λ_max is obtained. The equation is as follows: $P ω = λ_{max} ω$ (4)

After the weight distribution is normalized by the obtained feature vector ω, the importance ranking of each evaluation factor is obtained.

Consistency check. In order to check the correctness of the weights calculated above, this article performs a consistency check on the constructed matrix, and its formula is:

$CR = \frac{CI}{RI}$ (5)

In the formula, CR is the random consistency ratio of the judgment matrix, CI is the general consistency index of the judgment matrix, and RI is the evaluation random consistency index of the judgment matrix. Among them, the calculation formula of CI is: $CR = \frac{λ_{max} - n}{n - 1}$ (6)

The 1–9 order judgment matrix RI values are shown in Table 1 and Fig. 1. When the order is less than 2, we think that the matrix has complete consistency, and when the order is greater than 2, when CR < 0.10 or λ_max = n, CI = 0, we think that the P judgment matrix has satisfactory consistency. Otherwise, we need to adjust the values of the elements of the judgment matrix to recalculate the results until a satisfactory consistency is obtained.

Table 1

RI statistical table of average random consistency index

	RI	Order	RI
1	0	8	1.45
2	0	9	1.49
3	0.58	10	1.52
4	0.9	11	1.55
5	1.12	12	1.56
6	1.32

Fig. 1

RI statistics table of average random consistency index.

After passing the consistency test, this article can calculate the weight of each level of indicators.

2. Entropy weight method

The concept of entropy is a thermodynamic concept introduced by Shannon into information theory. Knowing from the basic principles of information theory, entropy is disordered in the system, and information is ordered, so the two signs are opposite. The basic principle is: the smaller the degree of disorder in the system, the smaller the entropy value, and the larger the amount of information, indicating that the index weight value is larger. Meanwhile, the greater the degree of disorder in the system, the greater the entropy value, the smaller the amount of information, and the smaller the role in evaluation research, indicating that the index weight value is smaller. Therefore, the entropy weight method can be used for comprehensive evaluation research. If multiple evaluation objects have the same calculated value in the same indicator, it indicates that the indicator does not play a role in the evaluation study. Therefore, this indicator should be removed from the indicator system [24].

The specific calculation steps are as follows:

(1) The original data matrix is constructed

If it is assumed that there are m objects to be evaluated and n evaluation index factors, the original data index matrix is formed as follows $X = {(x_{ij})}_{m \times n} = [\begin{matrix} x_{11} & x_{12} & \dots & x_{1 n} \\ x_{21} & x_{22} & \dots & x_{2 n} \\ ⋮ & ⋮ & \dots & ⋮ \\ x_{m 1} & x_{m 2} & \dots & x_{mn} \end{matrix}]$ (7)

(2) Normalization processing of the source data matrix

Before evaluation and analysis, the dimension source data of different dimensional index sources should be eliminated according to a standardized formula. The larger the standardized value, the better the evaluation result. After the matrix is normalized, the resulting matrix is as follows: $R = {(r_{ij})}_{m \times n} = [\begin{matrix} r_{11} & r_{12} & \dots & r_{1 n} \\ r_{21} & r_{22} & \dots & r_{2 n} \\ ⋮ & ⋮ & \dots & ⋮ \\ r_{m 1} & r_{m 2} & \dots & r_{mn} \end{matrix}]$ (8)

The normalized formula of the positive indicator is: $r_{ij} = \frac{x_{ij} - min_{j} {x_{ij}}}{max_{j} {x_{ij}} - min_{j} {x_{ij}}}$ (9)

The normalized formula of the negative indicator is: $r_{ij} = \frac{max_{j} {x_{ij}} - x_{ij}}{max_{j} {x_{ij}} - min_{j} {x_{ij}}}$ (10)

(3) Entropy is defined

According to the definition of entropy, the entropy of the jth index is defined as: $h_{j} = - k \sum_{i - 1}^{m} f_{ij} {Inf}_{ij} (i = 1, 2, \dots, m; j = 1, 2, \dots, n)$ (11)

Among them: $f_{ij} = \frac{r_{ij}}{\sum_{i - 1}^{m} r_{ij}}, k = \frac{1}{Inm}$ (12)

It is generally necessary to assume that at f_ij = 0, f_ijInf_ij = 0, which makes Inf_ij meaningful.

(4) The entropy weight is defined

The entropy weight of the jth indicator is: $W_{j} = \frac{1 - h_{j}}{n - \sum_{j = 1}^{n} h_{j}} (j = 1, 2, \dots, n)$ (13)

3. Combined weighting

This study combines the advantages of the subjective weighting method and the objective weighting method, and considers reducing the subjective arbitrariness and computational complexity of the decision-makers on the indicators. Moreover, this paper uses the combination weight method to determine the index weight to improve the scientific and reliability of the evaluation results.

This article first combines the index weight w₁ (j) calculated by the analytic hierarchy process with the index weight w₂ (j) calculated by the entropy weight method. After that, according to the principle of minimum relative information entropy, the subjective weight and the objective weight are combined to obtain the combined weight w (j) , (j = 1 ∼ n). The combined weight w (j) and w₁ (j) and w₂ (j) should be as close as possible. According to the principle of minimum relative information entropy, there are:

$\begin{matrix} min F = \sum_{j = 1}^{n} w (j) [Inw (j) - w_{1} (j)] \\ + \sum_{j = 1}^{n} w (j) [Inw (j) - w_{2} (j)] \end{matrix}$ (14)

Among them: $\sum_{j = 1}^{n} w (j) = 1, w (j) > 0, j = 1 \sim n$ (15)

When the Lagrangian multiplier method is used to solve the above optimization problem, the following results are obtained: $w (j) = \frac{{[w_{1} (j) w_{2} (j)]}^{0.5}}{\sum_{j = 1}^{m} {[w_{1} (j) w_{2} (j)]}^{0.5}} (j = 1, 2, \dots, n)$ (16)

The above formula shows that: among all the combined weights that satisfy formula (16), we can only obtain the least amount of information when we choose the geometric mean. However, when we choose other forms of combined weights, some additional information will appear invisible [25].

4 Establishment of evaluation model

Based on the established grassland resource distribution and utilization index system, this paper processes the index calculation results accordingly and synthesizes three first-level indexes: resource endowment index, resource distribution index, and resource utilization index. Finally, in this paper, the grassland resource distribution and utilization index is weighted by the results of the first-level index, and an evaluation model is established. The indicator system is mainly based on geographic census data, considering the distribution and utilization of grassland resources from different aspects. Therefore, it is realistic, highlights indicators that have an impact on grassland resources, and comprehensively evaluates the distribution and utilization of grassland resources and follows the principles of simple index calculation and strong maneuverability. Meanwhile, all indexes in this study are calculated using the comprehensive index method.

The comprehensive index method is a relatively simple quantitative evaluation method. When qualitatively and quantitatively analyzing each evaluation index factor, it uses formulas (17) to (20) for calculation according to the weight of each evaluation index factor. Finally, the grassland resource distribution and utilization index in each evaluation unit is obtained, and the calculation formula is as follows. ${Cov}_{1} = \sum_{i = 1}^{8} W_{i} p_{i}$ (17) ${Cov}_{2} = \sum_{i = 1}^{4} W_{i} p_{i}$ (18) ${Cov}_{3} = \sum_{i = 1}^{6} W_{i} p_{i}$ (19) $RDU = \sum_{i = 1}^{3} W_{i} {Cov}_{i}$ (20)

Among them: RDU (Resource distribution and utilization) is the grassland resource distribution and utilization index, p_i is the calculation result of the i-th evaluation index, and w_i is the weight value of the i-th evaluation index. The weight is determined by the combined weight method. At the same time, Cov_i (i = 1, 2, 3) is the calculation result of the i-th first-level index value, which refers to the resource endowment index, resource utilization index, and resource distribution index, respectively.

5 Graded assessment

Grassland resource distribution and utilization includes three dimensions: resource endowment, resource distribution, and resource utilization. In order to intuitively comprehensively analyze the grassland resource endowment, resource distribution, resource utilization and grassland resource comprehensive index scores in the evaluation unit and the relative differences between different evaluation units, and to evaluate the grassland resource distribution from different first-level indexes and overall perspectives situation and utilization status, this paper separately standardized the grassland resource distribution and utilization index to convert the index to [0, 100]. The specific calculation formula is as follows (taking the RDU index as an example): ${RDU}_{i}^{*} = \frac{{RDU}_{i}}{max ({RDU}_{i})} \times 100$ (21)

In the above formula: ${RDU}_{i}^{*}$ refers to the standardized score of the distribution and utilization index of grassland resources, max(RDU_i) refers to the maximum index value of the calculation result of the distribution and utilization of grassland resources in the i-th evaluation unit, and RDU_i refers to the value of grassland resource distribution and utilization index in the i-th evaluation unit.

Based on the above evaluation model principles, the evaluation and evaluation units can be calculated, and the grassland resource distribution and utilization index scores of each evaluation unit can be obtained. If the amount of data in the data sample is large, the data results cannot be seen intuitively. Therefore, on the basis of the above model, this article converts the quantitative index into qualitative analysis, that is, the index score is graded and evaluated. Moreover, for the three dimensions of resource endowment, resource distribution, and resource utilization, this paper formulates qualitative grade significance and divides grassland resource distribution and utilization into five levels: excellent, good, medium, poor, and poor. After that, this paper judges and analyzes the quality of grassland resources, the balance of grassland resource distribution and the degree of grassland resource utilization of the evaluation unit based on this grading standard.

Based on the above model formulas and calculation steps, this paper calculates and evaluates the unit to obtain the comprehensive index value of grassland resource distribution and utilization in each unit. Moreover, this paper ranks the index value after standardization to judge the quality, spatial distribution and utilization status of grassland resources in the region, so as to promote the reasonable distribution and sustainable utilization of grassland resources.

Based on the construction of the evaluation index system, this article uses a combination of subjective weighting method and objective weighting method to determine the index weight. In addition, this paper uses reasonable theoretical methods to construct a grassland resource distribution and utilization evaluation model, and introduces the principles of index standardization conversion and graded evaluation.

6 Model building

This study uses a three-layer BP network to build a red tourism resource evaluation model, and its network topology is shown in Fig. 2.

Fig. 2

BP neural network topology.

In this study, the information entropy weight method and extension method are introduced to evaluate and analyze the carrying capacity of grassland resources, and to analyze the differences between the administrative regions and theoretical analysis of sustainable development. According to matter-element extension theory, this paper builds a research and analysis system of water resources carrying capacity based on entropy weight extension decision theory. The flow chart of grassland resource carrying capacity evaluation is shown in Fig. 3.

Fig. 3

Flow chart of grassland resource carrying capacity evaluation.

This paper analyzes the different levels of resource and environmental carrying capacity to achieve the purpose of comprehensive evaluation of resource and environmental carrying capacity. Based on the theory of sustainable development and guided by the principles of index system construction, this paper considers the actual situation of the study and the availability and operability of the data, and combines with the opinions given by experts to finally form a specific index of grassland resources and environmental carrying capacity evaluation index system. Its structure is shown in Fig. 4.

Fig. 4

Structure diagram of bearing capacity index system.

This paper proposes an improved random forest regression model for the above problems. The improved random forest regression model has the advantages that the random forest model is not prone to overfitting, the training time is short, and the prediction effect is good. At the same time, it is more in line with the resource usage characteristics of cloud platform tasks to ensure the normal operation of tasks. The specific structure is shown in Fig. 5.

Fig. 5

The improved random forest regression model.

This article uses the resource usage data for the first 8 hours of the task to predict the maximum CPU utilization for the task in the 9th hour. The decision tree is shown in Fig. 6. The process of predicting the maximum CPU utilization for the 9th hour of a new task is: First, this article determines whether the maximum CPU usage for the 8th hour of the task exceeds 0.7. If it exceeds 0.7, it is judged that the maximum CPU usage rate of the task in the 9th hour is O.75. Otherwise, this article judges whether its priority is greater than 5. If it is greater than 5, the output value is 0.5. Otherwise, this article observes whether its average CPU usage in the eighth hour exceeds O.57. If the average CPU usage exceeds O.57, 0.4 is output, otherwise 0.2 is output.

Fig. 6

Decision tree model.

Based on the above analysis, this article uses actual cases to take research and analysis. According to the census data of the regional pastures in this study, the grassland area has also undergone certain changes over time. This paper analyzes the changes in the amount of grassland resources based on the changes in the amount of grassland resources from 2008 to 2019, as shown in Table 2 and Fig. 7.

Table 2

Changes in grassland area from 2008 to 2019 (unit: hectare)

	grassland	Natural grass	Improved grassland	Artificial grass
2008	223963.5	195284.7	3281.4	25397.4
2009	226296.3	195419.4	3281.4	27595.5
2010	227912.3	195413.1	3281.4	30124.9
2011	233101.9	195450.6	3281.4	34369.8
2012	236261.6	195449.9	3281.4	37530.3
2013	236257.6	195412.9	3281.4	37563.3
2014	237031.9	195429.6	3281.4	38320.9
2015	250765.6	195821.2	3281.4	51662.9
2016	253073.6	196502.7	3281.4	53289.5
2017	258254.8	196504.2	3281.4	53289.5
2018	268720.7	212477.2	3281.4	52962.1
2019	224716.6	157894.3	3281.4	63541.0

Fig. 7

Change trend of grassland.

As can be seen from the chart above, the area of pastureland is increasing from 2008 to 2019. The main reason for the increase is the increasing area of artificial grassland. Meanwhile, the changes are shown in Table 3 and Fig. 8 below.

Table 3

2017–2019 pasture area change table (unit: hectare)

	grassland	Natural grass	Artificial grass
2017	224716.6	157894.3	66821.5
2018	223911.3	157833.7	65067.5
2019	217216.8	149806.0	67410.6

Fig. 8

Changes in pasture area in 2017–2019 (unit: hectares).

The measurement of grassland resource quality is multifaceted. Among them, the grassland yield is an important aspect to measure the quality of grassland resources. According to the census data of Chayouzhongqi pasture, the grass yield of Chayouzhongqi pasture has shown a serious downward trend, which shows that the quality of pasture is declining continuously. The change of grassland yield in each time period is shown in Table 4 and Fig. 9 below.

Table 4

Change table of grassland yield in each time period (unit: hectare)

	Average annual grass yield of grassland (kg/mu)
The 60’s	228.9
The 80’s	103.4
The 90s	61.7
2007–2010	42.6
2011–2016	27.8
2017–2019	37.8

Fig. 9

Change graph of grassland yield in each time period (unit: hectare).

It can be clearly seen from the above chart that the grassland production has not exceeded this standard since 2014, which shows that the quality of grassland resources has been declining since 2014. The reason is that in recent years, the prices of livestock products have been increasing, and more and more people regard animal husbandry as the primary industry, and the number of livestock in Chayouzhongqi has grown rapidly, which eventually leads to overload and overgrazing. Moreover, successive years of drought have caused the carrying capacity of animals to exceed the carrying capacity of grasslands. In addition, the serious desertification has caused the grass production of the grassland to decrease year by year and the quality of the grassland to decline.

7 Conclusion

To achieve a good ecological environment, economic development, social stability, and sustainable useof grasslands, grassland resources should be used reasonably, and natural grasslands should be protected. The protection of natural grassland is mainly from the perspective of human factors, such as limiting overgrazing, which is the main factor of grassland degradation. In order to effectively control the serious degradation of natural pastures caused by overgrazing and protect natural grassland resources, it is necessary to adopt grazing prohibition measures in the region. Only when grazing is banned, plants can grow and develop normally, and grassland resources can be recovered, and this measure is conducive to better protection of grassland resources. At the same time, we must combine local economic and natural resource development needs to prevent the occurrence of overload and overgrazing. In addition, we need to adopt measures suiting local conditions and make the best use of the land, to resolve the contradiction between people’s immediate interests and long-term interests of grassland, so that people and nature can develop in harmony. Ultimately, we must achieve the coordination and unification of economic benefits, social benefits, and ecological benefits to achieve the sustainable use of grassland resources.

Footnotes

Acknowledgments

This project was funded by Jiangsu Province Science [BE2012340] and technology support project and National Natural Science Foundation of China [30972136].

References

Cai

, Luo

, Wang

, et al., Feature selection in machine learning: A new perspective, Neurocomputing 300(5) (2018), 70–79.

Goetz

J.N.

, Brenning

, Petschko

, et al., Evaluating machine learning and statistical prediction techniques for landslide susceptibility modeling, Computers & Geosciences 81(3) (2015), 1–11.

Darabi

, Choubin

, Rahmati

, et al., Urban flood risk mapping using the GARP and QUEST models: A comparative study of machine learning techniques, Journal of Hydrology 569(11) (2019), 142–154.

Rajkomar

, Dean

and Kohane

, Machine learning in medicine, New England Journal of Medicine 380(14) (2019), 1347–1358.

Xin

, Kong

, Liu

, et al., Machine learning and deep learning methods for cybersecurity, IEEE Access 6(3) (2018), 35365–35381.

Meredig

, Agrawal

, Kirklin

, et al., Combinatorial screening for new materials in unconstrained composition space with machine learning, Physical Review B 89(9) (2014), 094104.

Ward

, Agrawal

, Choudhary

, et al., A general-purpose machine learning framework for predicting properties of inorganic materials, Npj Computational Materials 2(1) (2016), 1–7.

Feng

, Wang

, Li Liu

, et al., Incorporating machine learning with biophysical model can improve the evaluation of climate extremes impacts on wheat yield in south-eastern Australia, Agricultural and Forest Meteorology 275(7) (2019), 100–113.

Kourou

, Exarchos

T.P.

, Exarchos

K.P.

, et al., Machine learning applications in cancer prognosis and prediction, Computational and Structural Biotechnology Journal 13(2) (2015), 8–17.

10.

Amershi

, Cakmak

, Knox

W.B.

, et al., Power to the people: The role of humans in interactive machine learning, Ai Magazine 35(4) (2014), 105–120.

11.

Rodriguez-Galiano

, Sanchez-Castillo

, Chica-Olmo

, et al., Machine learning predictive models for mineral prospectivity: An evaluation of neural networks, random forest, regression trees and support vector machines, Ore Geology Reviews 71(13) (2015), 804–818.

12.

Coley

C.W.

, Barzilay

, Jaakkola

T.S.

, et al., Prediction of organic reaction outcomes using machine learning, ACS Central Science 3(5) (2017), 434–443.

13.

Chowdhury

, Kautz

, Yener

, et al., Image driven machine learning methods for microstructure recognition, Computational Materials Science 123(22) (2016), 176–187.

14.

Basith

, Manavalan

, Shin

T.H.

, et al., SDM6A: A web-based integrative machine-learning framework for predicting 6mA sites in the rice genome, Molecular Therapy-Nucleic Acids 18(3) (2019), 131–141.

15.

Trost

S.G.

, Zheng

and Wong

W.K.

, Machine learning for activity recognition: hip versus wrist data, Physiological Measurement 35(11) (2014), 2183.

16.

Voyant

, Notton

, Kalogirou

, et al., Machine learning methods for solar radiation forecasting: A review, Renewable Energy 105(21) (2017), 569–582.

17.

Folberth

, Baklanov

, Balkovič

, et al., Spatio-temporal downscaling of gridded crop model yield estimates based on machine learning, Agricultural and Forest Meteorology 264(13) (2019), 1–15.

18.

Sieg

, Flachsenberg

and Rarey

, In need of bias control: evaluating chemical data for machine learning in structure-based virtual screening, Journal of Chemical Information and Modeling 59(3) (2019), 947–961.

19.

Thabtah

and Peebles

, A new machine learning model based on induction of rules for autism detection, Health Informatics Journal 26(1) (2020), 264–286.

20.

Narudin

F.A.

, Feizollah

, Anuar

N.B.

, et al., Evaluation of machine learning classifiers for mobile malware detection, Soft Computing 20(1) (2016), 343–357.

21.

Yao

, Yang

, Zhu

, et al., Core, mode, and spectrum assignment based on machine learning in space division multiplexing elastic optical networks, IEEE Access 6(3) (2018), 15898–15907.

22.

Bzdok

and Meyer-Lindenberg

, Machine learning for precision psychiatry: opportunities and challenges, Biological Psychiatry: Cognitive Neuroscience and Neuroimaging 3(3) (2018), 223–230.

23.

Chen

, Hao

, Hwang

, et al., Disease prediction by machine learning over big data from healthcare communities, Ieee Access 5 (2017), 8869–8879.

24.

Itu

, Rapaka

, Passerini

, et al., A machine-learning approach for computation of fractional flow reserve from coronary computed tomography, Journal of Applied Physiology 121(1) (2016), 42–52.

25.

Jayasinghe

, Lee

G.M.

, Um

T.W.

, et al., Machine learning based trust computational model for IoT services, IEEE Transactions on Sustainable Computing 4(1) (2018), 39–52.

26.

Rathore

M.M.

, Paul

, Hong

W.H.

, Seo

H.C.

, Awan

and Saeed

, Exploiting IoT and big data analytics: Defining smart digital city using real-time urban data, Sustainable Cities and Society 40 (2018), 600–610.

27.

Stefanova

, Hristova

and Golev

, Discrete Leader-Following Consensus for Multi-Agent System with Non-Instantaneous Impulses, Dynamic Systems and Applications 28 (2019), 743–755.

28.

Yimnet

and Niamsup

, Finite-Time Stability of Discrete Linear Singular Switched Positive Systems with Time-Delay by a Mode-Dependent Average Dwell Time Approach, Dynamic Systems and Applications 29(2) (2020), 192–219.

29.

Paul

, Internet of Things: A primer’, R Jeyaraj Human Behavior and Emerging Technologies 1(1) (2019), 37–47.

30.

Paul

, Daniel

, Ahmad

and Rho

, Cooperative cognitive intelligence for internet of vehicles, IEEE Systems Journal 11(3) (2017), 1249–1258.

31.

Paul

, Ahmad

, Mazhar Rathore

and Jabbar

, Smartbuddy: defining human behaviors using big data analytics in social internet of things, IEEE Wireless Communications 23(5) (2016), 68–74.

32.

Paul

, Jiang

Y.C.

, Wang

J.F.

and Yang

J.F.

, Parallel reconfigurable computing-based mapping algorithm for motion estimation in advanced video coding, ACM Transactions on Embedded Computing Systems (TECS) 11(S2) (2012), 1–18.