Multi-scene-driven low-carbon industrial park land utilization elastic planning method and system

Abstract

The invention provides a multi-scene-driven low-carbon industrial park land utilization elastic planning method and system. The multi-scene-driven low-carbon industrial park land utilization elastic planning method comprises the following steps of S1, data collection and preprocessing, S2, scene setting, S3, multi-scene simulation, S4, elastic index system construction, S5, scheme evaluation and optimization, and S6, planning scheme making. According to the invention, through multi-scene simulation, various possibilities of future development of the industrial park, including factors such as different economic development speeds, policy orientation and environmental changes, are fully considered, so that the planning scheme has adaptability and flexibility under various scenes. In the aspects of low-carbon development and decision support, a multi-attribute decision method based on an evidence theory, an improved particle swarm optimization algorithm and the like are adopted, so that a plurality of targets and indexes are comprehensively considered. Meanwhile, through a real-time monitoring and dynamic adjustment mechanism, adjustment and optimization are carried out according to actual conditions, and effective implementation of a planning scheme is ensured.

Description

Technical Field

[0001] The present invention relates to a multi-scenario driven low-carbon industrial park land use flexibility planning method, and also relates to a multi-scenario driven low-carbon industrial park land use flexibility planning system, belonging to the field of industrial planning software. Background Art

[0002] As global climate change becomes increasingly severe, developing a low-carbon economy has become a global consensus. Industrial parks, as key vehicles for industrial cluster development, contribute significantly to carbon emissions. Achieving low-carbon development in these parks is crucial for advancing regional and even global energy conservation and emission reduction goals. Traditional land use planning approaches for industrial parks primarily focus on functional zoning and spatial layout, often overlooking the requirements of low-carbon development and the uncertainties of future development.

[0003] In traditional planning, land use types are relatively fixed, lacking flexibility and adaptability. When industrial parks face shifts in market demand, technological innovation, or policy adjustments, rapid and effective land use adjustments are difficult to implement, leading to resource waste and increased carbon emissions. For example, some industrial parks fail to fully consider the development needs of emerging industries during planning. When emerging industries move in, the existing land use layout cannot meet their production and operational requirements, necessitating large-scale land transformation and reconstruction, which not only increases construction costs but also generates significant carbon emissions.

[0004] Furthermore, traditional planning methods have limitations when it comes to calculating and assessing carbon emissions. They typically employ static carbon emission factors, failing to account for dynamic changes in the energy structure, the impact of technological advancements on energy efficiency, and the phased nature of industrial development. This makes it difficult to accurately predict and control carbon emissions during implementation, hindering the effective achievement of low-carbon development goals. Furthermore, traditional planning methods lack the ability to simulate and analyze multiple scenarios, failing to fully account for potential future uncertainties such as economic fluctuations and environmental disasters, resulting in weak risk resilience.

[0005] To sum up, traditional land use planning methods for industrial parks can no longer meet the needs of low-carbon development and sustainable development. There is an urgent need for an innovative planning method that can comprehensively consider multiple scenario factors, improve the elasticity and adaptability of land use, effectively reduce carbon emissions, and achieve low-carbon, efficient and sustainable development of industrial parks. Summary of the Invention

[0006] The purpose of the present invention is to provide a multi-scenario driven low-carbon industrial park land use flexibility planning method and system to solve the above problems.

[0007] The present invention adopts the following technical solutions:

[0008] The present invention provides a multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method, which is characterized by comprising the following steps:

[0009] S1: Data collection and preprocessing steps: Receive basic data of industrial parks collected by IoT sensor networks, satellite remote sensing, and drone aerial surveys, use deep learning anomaly data detection algorithms to clean the data, and then normalize the data. The formula is x is the original data, μ(x) is the mean, σ(x) is the standard deviation, and ∈ is the minimum value to prevent the denominator from being zero;

[0010] S2: Scenario setting step: Use the conditional generative adversarial network to input different factors as conditions to generate land use scenarios, determine the key driving factors and their value ranges, and use the combination of the analytic hierarchy process (AHP) and the entropy weight method (EWM) to determine the weights of each driving factor. The formula is: Where α is the weight adjustment coefficient, are the weights of the i-th driving factor calculated by AHP and EWM respectively;

[0011] S3: Multi-scenario simulation steps: Construct a land use simulation model based on the improved cellular automation-multi-agent system (CA-MAS model) coupled with system dynamics, simulate the dynamic feedback of multiple systems in the industrial park, and calculate the total carbon emissions using the carbon emission dynamic accounting model. The formula is Among them C t is the total carbon emissions, E i,t is the energy consumption, F i,t is the energy carbon emission factor, β i，t is the correction factor;

[0012] S4: Steps for constructing a resilience indicator system: Establish a dynamic land use resilience indicator system that includes structure, function, time, and resilience. The structural resilience indicator uses the improved Shannon diversity index. Among them, P i is the area proportion of land use type, is the area proportion of land use types under ideal conditions;

[0013] S5: Scheme evaluation and optimization step: Use the multi-attribute decision-making method based on evidence theory to evaluate the scheme, use the improved particle swarm optimization algorithm to optimize the scheme, and analyze the sensitivity and robustness of the scheme;

[0014] S6: Steps for formulating planning schemes: Adopt a combination of dynamic and rolling planning methods, divide the planning schemes into short-, medium- and long-term goals, clarify the spatial layout and functional positioning of various types of land, and formulate ecological compensation mechanisms and low-carbon incentive policies.

[0015] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0016] Step S1 also includes the step of collecting stakeholder data. The information collection module receives the data submitted by stakeholders, votes on the plan, and automatically processes the voting results according to preset rules.

[0017] In step S3, there is also a step of using a graph neural network to model the relationship between cells and agents.

[0018] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0019] The carbon emission accounting model in step S3 also has carbon sink parameters, and the corresponding carbon emission dynamic accounting model formula is: Among them S t is the total carbon sink in the industrial park at time t.

[0020] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0021] In S4: the step of constructing a resilience indicator system, the resilience indicator includes an ecological resilience indicator, which includes an ecological footprint indicator and an ecological risk indicator. The evaluation is performed by combining a fuzzy comprehensive evaluation method with a risk matrix method.

[0022] Step S4 also includes the step of dynamic early warning of ecological resilience indicators:

[0023] When the ecological resilience index is lower than the set warning threshold, a warning signal is issued.

[0024] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0025] In S5: solution evaluation and optimization step, a multi-objective evolutionary algorithm is also included:

[0026] The non-dominated sorting genetic algorithm II is used to iteratively search and obtain the Pareto optimal solution set. The preference information entropy method is used to screen and sort the Pareto optimal solution set, and multiple candidate decision plans are obtained. At the same time, combined with the spatial analysis function of the geographic information system, the candidate decision plans are spatially visualized and analyzed.

[0027] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0028] In S6: Planning scheme formulation step, a dynamic land use zoning strategy based on scenario perception is generated. The scenario perception technology of artificial intelligence is used to perceive the development scenario and environmental changes of the industrial park in real time. According to different scenarios, the industrial park is divided into different dynamic zones. Different land use functions and development strategies are generated for each zone, and corresponding differentiated land use policies and management measures are generated.

[0029] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0030] In the step S1: Data collection and preprocessing, edge computing technology is used to locally process and analyze the collected data. Edge computing equipment is deployed at the data collection site to perform preliminary processing of the data in real time. At the same time, a federated learning algorithm is used to collaboratively analyze multi-source data to achieve knowledge sharing and model training between different data sources.

[0031] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0032] In S2: scenario setting step, there is also a policy model simulation step:

[0033] Based on a method that combines system dynamics and machine learning, the development trends of industrial parks under different policy scenarios are simulated. The policy test bed method is used to test and evaluate different policy combinations, analyze the implementation effects and potential impacts of policies, obtain policy simulation results, and use the policy simulation results as input conditions to generate land use scenarios.

[0034] Furthermore, the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method of the present invention also has the following characteristics:

[0035] After step S6, there is also step S7: real-time evaluation step:

[0036] During the implementation of the planning scheme, a performance evaluation and incentive mechanism based on blockchain will be established. The distributed ledger technology of blockchain will be used to record the implementation process and performance data of the planning scheme. The balanced scorecard method will be used to build a performance evaluation indicator system. The implementation effect of the planning scheme will be evaluated from four dimensions: finance, customers, internal processes, and learning and growth. Based on the evaluation results, incentive rewards will be automatically issued through smart contracts. At the same time, big data analysis technology will be used to conduct in-depth mining and analysis of performance evaluation data to provide decision support for the adjustment and optimization of the planning scheme.

[0037] After step S6, it also includes real-time monitoring and dynamic adjustment steps. An intelligent monitoring network is deployed in the industrial park. The monitoring equipment realizes real-time transmission and sharing of data through 5G communication technology. Blockchain technology is used to build a data trusted storage and sharing platform. The real-time monitoring data is input into the online simulation model. When the deviation between the actual situation and the expected scenario exceeds the set threshold, a method based on rule reasoning and case reasoning is used to make dynamic adjustments. Rule reasoning makes decisions based on a preset rule library, and case reasoning searches for similar cases from the historical case library for reference.

[0038] Using transfer learning technology, models trained in other similar industrial parks are migrated to the current industrial park.

[0039] This embodiment also provides a multi-scenario simulation-driven low-carbon industrial park flexible planning system, including:

[0040] The data collection and preprocessing module receives basic industrial park data collected by IoT sensor networks, satellite remote sensing, and drone aerial surveys, uses deep learning anomaly data detection algorithms to clean the data, and then normalizes the data;

[0041] The scenario setting module runs a conditional generative adversarial network, takes different factors as input, generates land use scenarios, determines key driving factors and their value ranges, and uses the combination of the analytic hierarchy process (AHP) and the entropy weight method (EWM) to determine the weights of each driving factor.

[0042] The multi-scenario simulation module builds a land use simulation model based on the improved cellular automation-multi-agent system CA-MAS model coupled with system dynamics, simulates the dynamic feedback of multiple systems in the industrial park, and inputs it into the carbon emission dynamic accounting model to obtain the total carbon emissions of the park;

[0043] The elasticity indicator system construction module is used to establish a dynamic land use elasticity indicator system including structure, function, time and resilience.

[0044] The scheme evaluation and optimization module uses a multi-attribute decision-making method based on evidence theory to evaluate schemes, an improved particle swarm optimization algorithm to optimize schemes, and analyzes the sensitivity and robustness of schemes;

[0045] The planning scheme formulation module adopts a combination of dynamic and rolling planning methods to divide the planning scheme into short-, medium- and long-term goals, clarify the spatial layout and functional positioning of various types of land, and generate ecological compensation mechanisms and low-carbon incentive policies;

[0046] The feedback data collection module uses the data collection and preprocessing module to continue to receive implementation process and performance data during the execution of the planning scheme, and adopts the balanced scorecard method to build a performance evaluation indicator system to evaluate the implementation effect of the planning scheme from four dimensions: finance, customers, internal processes, and learning and growth.

[0047] Beneficial Effects of the Invention: In terms of the scientific and adaptable nature of planning, this invention, through multi-scenario simulation, fully considers the various possibilities for the future development of industrial parks, including factors such as varying economic development rates, policy orientations, and environmental changes. This allows planning schemes to be no longer single and static, but rather adaptable and flexible across multiple scenarios. For example, in a high-growth, low-carbon scenario, the planning scheme can rationally allocate land, ensuring space for the development of emerging industries while effectively controlling carbon emissions. In a transitional development scenario, it can guide the smooth transformation of industrial structure and achieve optimal allocation of land resources.

[0048] In terms of low-carbon development, this invention introduces an advanced carbon emission accounting model that can accurately calculate and assess carbon emissions under different scenarios. By optimizing land use layout and industrial configuration, promoting efficient energy utilization and the promotion of clean energy, it effectively reduces the total carbon emissions of industrial parks. Furthermore, the established ecological resilience indicator system helps protect and enhance the ecological environment of industrial parks, strengthen carbon sequestration capacity, and further promote the achievement of low-carbon development goals.

[0049] In terms of decision support, this invention utilizes a variety of advanced evaluation and optimization algorithms, such as a multi-attribute decision-making method based on evidence theory and an improved particle swarm optimization algorithm. These algorithms comprehensively consider multiple objectives and indicators, providing decision makers with a scientific and objective basis for decision-making. Furthermore, through real-time monitoring and dynamic adjustment mechanisms, problems during the implementation of planning schemes can be promptly identified, and adjustments and optimizations can be made based on actual conditions to ensure the effective implementation of planning schemes.

[0050] In terms of stakeholder engagement, the stakeholder engagement platform built on blockchain and smart contracts has increased stakeholder participation and satisfaction. All parties can participate in the planning process in a transparent and fair environment, expressing their opinions and needs, making the planning plan more in line with actual conditions and the interests of all parties. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a schematic diagram of the multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method proposed in the present invention;

[0052] Figure 2 A bar chart comparing the adaptability scores of planning schemes under different planning methods;

[0053] Figure 3A line chart comparing changes in total carbon emissions during the planning period using the traditional method and this method;

[0054] Figure 4 The radar chart is a comparison of the economic benefit growth rates of the traditional method and this method. DETAILED DESCRIPTION

[0055] The following describes the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0056] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0057] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The present invention will be further described in detail below with reference to the accompanying drawings.

[0058] The multi-scenario simulation-driven low-carbon industrial park flexible planning system in this embodiment includes:

[0059] The data collection and preprocessing module receives basic industrial park data collected by IoT sensor networks, satellite remote sensing, and drone aerial surveys, uses deep learning anomaly data detection algorithms to clean the data, and then normalizes the data;

[0060] The scenario setting module runs a conditional generative adversarial network, takes different factors as input, generates land use scenarios, determines key driving factors and their value ranges, and uses the combination of the analytic hierarchy process (AHP) and the entropy weight method (EWM) to determine the weights of each driving factor.

[0061] The multi-scenario simulation module builds a land use simulation model based on the improved cellular automation-multi-agent system CA-MAS model coupled with system dynamics, simulates the dynamic feedback of multiple systems in the industrial park, and inputs it into the carbon emission dynamic accounting model to obtain the total carbon emissions of the park;

[0062] The elasticity indicator system construction module is used to establish a dynamic land use elasticity indicator system including structure, function, time and resilience.

[0063] The scheme evaluation and optimization module uses a multi-attribute decision-making method based on evidence theory to evaluate schemes, an improved particle swarm optimization algorithm to optimize schemes, and analyzes the sensitivity and robustness of schemes;

[0064] The planning scheme formulation module adopts a combination of dynamic and rolling planning methods to divide the planning scheme into short-, medium- and long-term goals, clarify the spatial layout and functional positioning of various types of land, and generate ecological compensation mechanisms and low-carbon incentive policies;

[0065] The feedback data collection module uses the data collection and preprocessing module to continue to receive implementation process and performance data during the execution of the planning scheme, and adopts the balanced scorecard method to build a performance evaluation indicator system to evaluate the implementation effect of the planning scheme from four dimensions: finance, customers, internal processes, and learning and growth.

[0066] Among them, the term resilience mentioned above refers to land use resilience (LUR), which is defined as the ability of the land system to maintain core functions, adapt to changes and achieve transformation when facing internal and external risk shocks.

[0067] Land use resilience is usually characterized by indicators in multiple dimensions, such as resource resilience indicators, ecological resilience indicators, economic resilience indicators, structural and morphological resilience indicators, and functional resilience indicators.

[0068] Reference Figures 1 to 4 The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method in this embodiment includes the following steps:

[0069] Step S1, Data Collection and Preprocessing: Comprehensively collect basic data from the industrial park using a combination of multi-source data collection methods, including IoT sensor networks, satellite remote sensing, and drone aerial surveys. IoT sensor networks will be deployed extensively at key locations within the industrial park, such as factory floors, energy supply stations, and public buildings. These sensors can monitor energy consumption in real time, such as electricity, gas, and water usage. They also collect environmental quality indicators, such as air pollutant concentrations, temperature, and humidity. This allows us to obtain dynamic data on energy and the environment within the industrial park. Satellite remote sensing technology is used to acquire large-scale geospatial data, providing high-resolution imagery that helps us understand the topography, land use, and other aspects of the industrial park. Drone aerial surveys, on the other hand, provide more detailed data collection in key areas, such as newly planned construction areas or ecologically sensitive areas, to obtain high-precision geospatial data. After the collected data is transmitted to the data center, we will utilize advanced data cleaning and preprocessing technologies. An abnormal data detection algorithm based on deep learning will be applied. Specifically, the isolation forest algorithm will be used to perform a preliminary screening of the data. This algorithm can quickly identify data points that are significantly deviated from the normal range and mark them as potential outliers. Then, the long short-term memory network (LSTM) will be used to further analyze the data. LSTM is a recursive neural network that can process time series data. By learning the time series characteristics and normal fluctuation range of the data, it can identify abnormal data hidden in the data. Once abnormal data is found, it will be corrected or eliminated to ensure data quality. Next, the cleaned data is normalized. The adaptive normalization method is used, and the formula is Where x is the original data, μ(x) is the mean of the data, σ(x) is the standard deviation of the data, and ∈ is a minimum value, usually set to 10 -6 , which is used to prevent the denominator from being zero. This adaptive normalization method dynamically adjusts the normalization parameters based on real-time data changes, ensuring that data from different sources are comparable.

[0070] Finally, a data feature extraction model was constructed, and the data was processed using a convolutional neural network (CNN). CNNs possess powerful feature extraction capabilities. Through multi-layer convolution and pooling operations, they can extract deep-level feature information from the data. For example, for geospatial data, CNNs can extract features such as the boundaries and textures of land use types; for energy consumption data, they can extract patterns and trends in energy use. These features provide high-quality data support for subsequent scenario setting and simulations.

[0071] Step S2, scenario setting: A scenario generator model is used to generate diverse land use scenarios based on a conditional generative adversarial network (CGAN). First, the input conditions are determined, which include the development goals, policy orientations, and external environmental factors of the industrial park. Development goals can be economic growth goals, low-carbon development goals, etc.; policy orientations include tax incentives, industrial access policies, etc.; external environmental factors include changes in market demand, the speed of technological innovation, etc. These conditions are encoded and input into the generator of CGAN. The generator will try to generate land use scenario data that meets the requirements based on the input conditions, while the discriminator will receive real land use data and data generated by the generator, and judge the authenticity of the data. Through adversarial training between the generator and the discriminator, the performance of the generator is continuously optimized, enabling it to generate more realistic and diverse land use scenarios.

[0072] For each generated scenario, we need to determine the key driving factors and their value ranges. Key driving factors may include industry growth rate, clean energy utilization rate, policy implementation strength, etc. In order to accurately determine the weight of each driving factor, we will use a combination of the analytic hierarchy process and entropy weight method (AHP-EWM). First, through the analytic hierarchy process (AHP), organize experts to compare each driving factor pairwise and construct a judgment matrix. The judgment matrix reflects the subjective judgment of experts on the relative importance of each driving factor. By calculating the eigenvector of the judgment matrix and the consistency test, the subjective weight of each driving factor is obtained. At the same time, the entropy weight method (EWM) is used to calculate the objective weight of each driving factor The entropy weight method determines the weight according to the degree of dispersion of the data. The greater the degree of dispersion of the data, the greater the influence of the driving factor on the result, and the higher its weight. Finally, the subjective weight and the objective weight are weighted and combined to obtain the comprehensive weight. Where α is the weight adjustment coefficient, which is set according to actual conditions and generally takes a value between 0.5 and 0.7 to balance the influence of subjective and objective factors.

[0073] Step S3, multi-scenario simulation: A land use simulation model based on an improved cellular automation-multi-agent system (CA-MAS) coupled with system dynamics (SD) is constructed. In CA-MAS, the industrial park is divided into multiple cells, each representing a land unit of a certain area. Cells have multiple states, including land use type (such as industrial land, commercial land, residential land, green space, etc.), building density, and energy consumption intensity. The state of a cell changes dynamically based on its own state and the states of neighboring cells. Using an adaptive rule adjustment mechanism, a cell dynamically adjusts its conversion rules based on the current scenario and its own state. For example, when industrial land in the neighborhood increases and traffic pressure increases, the probability of a cell converting to commercial land is adjusted accordingly according to the pre-set rules. This adaptive mechanism enables the model to better adapt to land use changes under different scenarios. The agents in the multi-agent system include different stakeholders such as enterprises, governments, and residents. Each agent has specific decision-making capabilities and behavioral rules. Enterprise agents use deep reinforcement learning algorithms (such as Deep Q Networks, or DQNs) to make land use decisions based on their development needs, cost-benefit analysis, and market conditions. For example, they consider whether to expand production or implement low-carbon technology transformation to maximize their own interests.

[0074] The government agent will formulate land use policies and regulatory measures based on policy objectives and regional development plans. Through tax policies, land approval policies, and other means, the government will guide the rational allocation of land resources and promote the low-carbon development of industrial parks.

[0075] Residents’ intelligent bodies will give feedback on changes in land use based on quality of life and environmental needs. For example, residents may raise concerns about noise pollution, air quality, and other issues, which will affect land use decisions. The system dynamics (SD) model is combined with CA-MAS to simulate the dynamic feedback relationship between the economic, environmental, and social systems in the industrial park. The SD model can describe the causal relationship and feedback mechanism between the variables in the system, and simulate the dynamic changes of the system by establishing a series of equations and feedback loops. For example, while industrial development drives economic growth, it will increase energy consumption and carbon emissions, thereby affecting environmental quality. Changes in environmental quality will in turn affect the quality of life of residents and the attractiveness of enterprises, thereby having a counter-effect on industrial development. The simulation of this dynamic feedback relationship enables us to have a more comprehensive understanding of the development process and potential problems of the industrial park. In terms of carbon emission accounting, the carbon emission dynamic accounting model is adopted, and the formula is Among them C t is the total carbon emissions at time t, E i,t is the consumption of the i-th energy at time t, F i，t is the carbon emission factor of the i-th energy source at time t, β i，tIt is a correction factor to take into account factors such as the improvement of energy efficiency and the application of carbon capture technology. i，t The carbon emission factor F can be collected in real time through IoT sensors. i，t It will be updated regularly based on the energy market and technological progress, and the correction coefficient β i，t It is necessary to comprehensively consider the improvement of energy efficiency (such as equipment updates, technological improvements, etc.) and the application effect of carbon capture technology, and conduct assessments and adjustments every quarter.

[0076] Step S4, elasticity index system construction: establish a dynamic land use elasticity index system, including structural elasticity, functional elasticity, temporal elasticity and resilience elasticity indicators. The structural elasticity index adopts the improved Shannon diversity index. Among them, P i is the area proportion of the i-th land use type, = is the ideal percentage of land use type i, pre-determined based on the industrial park's development positioning and ecological requirements. The improved Shannon Diversity Index more accurately reflects the diversity and rationality of land use types. A higher index value indicates a more rational and resilient land use structure.

[0077] Functional flexibility measures the degree of synergy between different land use functions using a functional coupling index. This evaluation model is constructed by analyzing the flow of materials, energy, and information between different land use functions. For example, the volume of cargo transported and the frequency of information exchange between industrial and logistics land are calculated to assess the efficiency of their synergy. A higher degree of functional coupling indicates greater synergy between land use functions, and thus greater adaptability to the development and changes of the industrial park.

[0078] The temporal elasticity indicator uses time series analysis to analyze and forecast historical land use data. Specifically, we use the ARIMA model (Autoregressive Integrated Moving Average) to fit and forecast land use data. By analyzing the changing trends and patterns of land use at different time scales, we assess the temporal elasticity of land use. For example, we can predict the area changes of different land use types over the next few years to determine whether land use can adapt to changes in industrial development and market demand.

[0079] The resilience index constructs a land use resilience network model based on complex network theory. Land use units are considered network nodes, and the relationships between units (such as transportation connections, industrial linkages, and ecological dependencies) are considered network edges. By calculating indicators such as network connectivity, clustering coefficient, and node importance, the resilience of industrial parks' land use systems in the face of sudden disasters (such as natural disasters and public health incidents) is assessed. Higher connectivity and a larger clustering coefficient indicate a more resilient land use system, meaning it can recover more quickly after damage.

[0080] Step S5, scheme evaluation and optimization: In the scheme evaluation and optimization stage, we will use the multi-attribute decision-making method based on evidence theory to evaluate the land use schemes obtained from different scenario simulations. First, determine the evaluation index set, including various elasticity indicators (structural elasticity, functional elasticity, time elasticity, resilience elasticity), total carbon emissions, economic benefit indicators (such as industrial added value, tax revenue, etc.), and social performance indicators (such as the number of jobs added, resident satisfaction, etc.). Then, based on expert scores and historical data, construct a trust function and likelihood function for each indicator. The trust function represents the degree of trust in a certain scheme on a certain indicator, and the likelihood function represents the maximum degree of trust that the scheme can achieve on the indicator. By calculating the trust of each scheme under different indicators, the comprehensive evaluation results of the scheme are obtained. The improved particle swarm optimization algorithm (PSO) is used to optimize the scheme. On the basis of the traditional PSO algorithm, adaptive inertia weight and mutation operator are used. The adaptive inertia weight w is dynamically adjusted according to the number of iterations and the distribution of particles. The formula is where w max and w min are the maximum and minimum values ​​of the inertia weight, t is the current iteration number, and T is the total iteration number. The mutation operator mutates particles with a certain probability to enhance the algorithm's global search capability and its ability to escape local optimality. After optimization, the sensitivity and robustness of the solution are analyzed by changing key parameters (such as industry growth rate, policy implementation strength) and uncertainty factors (such as market fluctuation range, technological innovation speed). The fluctuation of the solution's indicators under different parameter and factor changes is calculated to evaluate the solution's stability and adaptability. If the various indicators of the solution can still remain within an acceptable range when the parameters and factors of the solution change significantly, it means that the solution has strong robustness.

[0081] Step S6, planning scheme formulation: Based on the optimized scheme, a land use planning scheme for the industrial park is formulated by combining dynamic planning and rolling planning.

[0082] The planning period is divided into short-term (1-3 years), medium-term (3-5 years) and long-term (5-10 years), and the development goals and key tasks for each stage are clearly defined.

[0083] The short-term goal focuses on solving the current problems in industrial parks, such as optimizing land use layout and improving energy utilization efficiency; the medium-term goal focuses on industrial upgrading and structural adjustment, and promoting the development of low-carbon industries; the long-term goal aims to achieve the sustainable development of industrial parks and build an eco-friendly land use model.

[0084] In terms of spatial layout, different functional areas are divided, such as core development areas, flexible expansion areas, ecological protection buffer zones, etc.

[0085] Core development zones are key areas for industrial development, focusing on high-end manufacturing, R&D and innovation centers, and more. Flexible expansion zones are flexibly adjusted based on actual industrial development, providing space for the development of emerging industries. Ecological protection buffer zones strictly restrict development to ensure ecological and environmental safety, such as the construction of wetland reserves and forest parks. Land use regulations and control measures are formulated for different functional areas. For example, in core development zones, land use efficiency and environmental protection requirements are high, encouraging enterprises to adopt advanced production technologies and energy-saving and emission-reduction measures. Flexible expansion zones offer certain policy incentives to attract emerging industries. In ecological protection buffer zones, development activities that damage the ecological environment are strictly prohibited.

[0086] In developing an ecological compensation mechanism and low-carbon incentive policies, the ecological compensation mechanism provides differentiated compensation based on the degree of ecological impact of different land use types. Land users who contribute significantly to ecological protection, such as managers of ecological protection areas and green agricultural producers, will receive financial subsidies, tax exemptions, and other forms of compensation. Land users who cause ecological damage will be subject to ecological compensation fees. Low-carbon incentive policies include tax incentives, financial subsidies, and priority land use approvals for enterprises that adopt low-carbon technologies and green production methods, encouraging them to actively participate in low-carbon development.

[0087] In a preferred embodiment, real-time monitoring and dynamic adjustment steps are also included. A comprehensive intelligent monitoring network is deployed in the industrial park, which includes not only traditional environmental and energy monitoring equipment, but also intelligent traffic monitoring systems, intelligent building monitoring systems, etc. These monitoring devices realize real-time data transmission and sharing through 5G communication technology. Blockchain technology is used to build a trusted data storage and sharing platform to ensure the authenticity and non-tampering of monitoring data. The real-time monitoring data is input into the online simulation model, which is based on a real-time learning algorithm and can continuously update the model parameters and simulation results according to new data. When the deviation between the actual situation and the expected scenario exceeds the set threshold, a method based on a combination of rule reasoning and case reasoning is used for dynamic adjustment. Rule reasoning makes decisions based on a preset rule base, while case reasoning searches for similar cases from the historical case base for reference and reference, quickly generates adjustment plans and implements them.

[0088] In a preferred embodiment, the process also includes a stakeholder engagement step. A stakeholder engagement platform based on blockchain and smart contracts is constructed, ensuring transparency and traceability of information. During key stages such as scenario setting, scenario evaluation, and planning proposal development, relevant information is disseminated through the platform, inviting participation from stakeholders such as government departments, business representatives, and residents. Stakeholders can submit opinions and suggestions through the platform and vote on proposals. Smart contracts automatically process voting results according to pre-set rules and provide feedback to the planning team. Social network analysis is also employed to analyze the relationships and influence among stakeholders. Targeted communication and coordination are conducted based on these findings, enhancing stakeholder engagement and satisfaction.

[0089] In a preferred embodiment, the CA-MAS model undergoes a profound improvement during the multi-scenario simulation step, employing a graph neural network (GNN) to model the complex relationships between cells and agents. GNNs can capture the spatial topology and interactions between cells and agents, improving the model's simulation accuracy. Furthermore, the policy gradient algorithm employed in reinforcement learning is combined to optimize the agent's decision-making strategy. Through interaction with the environment, the agent continuously learns the optimal decision-making strategy to maximize land use efficiency.

[0090] In the carbon emission accounting model, the role of carbon sinks is further considered and carbon sinks are included in the accounting formula. The formula is improved to Among them S t is the total carbon sink in the industrial park at time t, including vegetation carbon sequestration, wetland carbon sequestration, etc.

[0091] Figure 2 、 Figure 3 and Figure 4 The advantages of the method provided in this embodiment compared with the traditional method are shown. Figure 2 It shows that the scoring of the present invention is better than that of the traditional method under the conditions of various types of parks. Figure 3 It is shown that the present invention has a significant reduction in total carbon emissions compared with traditional methods. Figure 4 It shows that the present invention has obvious advantages over traditional methods in five dimensions, including tax growth rate, economic benefit growth rate, return on investment, employment growth rate and industrial added value growth rate.

[0092] In the present invention, the ecological resilience index is further improved during the step of constructing the resilience index system. In addition to the ecosystem service value and biodiversity protection, the ecological footprint index and ecological risk index are added. The ecological footprint index measures the degree to which human activities in the industrial park occupy the ecological environment and is calculated using the ecological footprint model; the ecological risk index assesses the ecological risks faced by the industrial park, such as soil pollution risk and water resource shortage risk, using a combination of fuzzy comprehensive evaluation and risk matrix methods. At the same time, a dynamic early warning mechanism for the ecological resilience index is established. When the ecological resilience index falls below the set early warning threshold, an early warning signal is issued in a timely manner, and corresponding ecological restoration and protection measures are initiated.

[0093] In the present invention, in the scheme evaluation and optimization step, the multi-objective evolutionary algorithm MOEA is used to optimize the scheme. MOEA can simultaneously optimize multiple conflicting objectives, such as minimizing carbon emissions, maximizing land use efficiency, maximizing economic benefits, etc. The non-dominated sorting genetic algorithm II (NSGA-II) is used as the specific implementation algorithm of MOEA, and a set of non-dominated solutions, i.e., the Pareto optimal solution set, is obtained through iterative search. The preference information entropy method is used to screen and sort the Pareto optimal solution set, and the optimal scheme is selected according to the decision maker's preference. At the same time, combined with the spatial analysis function of the geographic information system (GIS), the optimized scheme is spatially visualized and analyzed to intuitively display the advantages and disadvantages and feasibility of the scheme.

[0094] In the present invention, a dynamic land use zoning strategy based on scenario perception is formulated during the planning scheme formulation step. Using artificial intelligence scenario perception technology, the development scenario and environmental changes of the industrial park are perceived in real time. Based on different scenarios, the industrial park is divided into different dynamic zones, such as innovation development zones, green transformation zones, and ecological conservation zones. Each zone has different land use functions and development strategies, and the boundaries and scope of the zones can be dynamically adjusted based on scenario changes. Differentiated land use policies and management measures are formulated for different dynamic zones to improve the flexibility and adaptability of land use.

[0095] In this invention, edge computing technology is used to locally process and analyze collected data during the data collection and preprocessing steps. Edge computing equipment, deployed at the data collection site, enables real-time preliminary data processing, reducing data transmission volume and latency. Furthermore, federated learning algorithms are used to collaboratively analyze multi-source data, enabling knowledge sharing and model training across different data sources while ensuring data privacy. Using transfer learning technology, models trained in other similar industrial parks are migrated to the current industrial park, improving model training efficiency and accuracy.

[0096] In this invention, the impact of dynamic policy changes is considered during the scenario setting step. A policy simulation model is constructed, based on a combination of system dynamics and machine learning, to simulate the development trends of industrial parks under different policy scenarios. A policy testbed approach is used to test and evaluate different policy combinations, analyzing the implementation effects and potential impacts of the policies. During the scenario generation process, the policy simulation results are used as input to generate land use scenarios that better align with the actual policy environment, thereby improving the policy adaptability of the planning scheme.

[0097] In this invention, a blockchain-based performance evaluation and incentive mechanism is established during the implementation of the planning scheme. The distributed ledger technology of the blockchain is used to record the implementation process and performance data of the planning scheme, ensuring the authenticity and traceability of the data. A performance evaluation indicator system is constructed using the balanced scorecard approach, evaluating the implementation effectiveness of the planning scheme from four dimensions: finance, customers, internal processes, and learning and growth. Based on the evaluation results, incentives are automatically issued through smart contracts to encourage relevant departments and enterprises to actively participate in the implementation of the planning scheme. At the same time, big data analysis technology is used to deeply mine and analyze performance evaluation data, providing decision support for the adjustment and optimization of the planning scheme.

[0098] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A multi-scenario simulation-driven land use flexibility planning method for low-carbon industrial parks, characterized by: The following steps are involved: S1: Data collection and preprocessing steps: Receive basic data of industrial parks collected by IoT sensor networks, satellite remote sensing, and drone aerial surveys, use deep learning anomaly data detection algorithms to clean the data, and then normalize the data. The formula is x is the original data, μ(x) is the mean, σ(x) is the standard deviation, and ∈ is the minimum value to prevent the denominator from being zero; S2: Scenario setting step: Use the conditional generative adversarial network to input different factors as conditions to generate land use scenarios, determine the key driving factors and their value ranges, and use the combination of the analytic hierarchy process (AHP) and the entropy weight method (EWM) to determine the weights of each driving factor. The formula is: Where α is the weight adjustment coefficient, are the weights of the i-th driving factor calculated by AHP and EWM respectively; S3: Multi-scenario simulation steps: Construct a land use simulation model based on the improved cellular automation-multi-agent system coupled with system dynamics, simulate the dynamic feedback of multiple systems in the industrial park, and calculate the total carbon emissions using the carbon emission dynamic accounting model. The formula is Among them C t is the total carbon emissions, E i，t is the energy consumption, F i，t is the energy carbon emission factor, β i,t is the correction factor; S4: Steps for constructing a resilience indicator system: Establish a dynamic land use resilience indicator system that includes structure, function, time, and resilience. The structural resilience indicator uses the improved Shannon diversity index. Among them, P i is the area proportion of land use type, is the area proportion of land use types under ideal conditions; S5: Scheme evaluation and optimization step: Use the multi-attribute decision-making method based on evidence theory to evaluate the scheme, use the improved particle swarm optimization algorithm to optimize the scheme, and analyze the sensitivity and robustness of the scheme; S6: Steps for formulating planning schemes: Adopt a combination of dynamic and rolling planning methods, divide the planning schemes into short-, medium- and long-term goals, clarify the spatial layout and functional positioning of various types of land, and formulate ecological compensation mechanisms and low-carbon incentive policies.

2. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: Also includes: Step S1 also includes the step of collecting stakeholder data. The information collection module receives the data submitted by stakeholders, votes on the plan, and automatically processes the voting results according to preset rules. In step S3, there is also a step of using a graph neural network to model the relationship between cells and agents.

3. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: The carbon emission accounting model in step S3 also has carbon sink parameters, and the corresponding carbon emission dynamic accounting model formula is: Among them S t is the total carbon sink in the industrial park at time t.

4. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized by: In S4: the step of constructing a resilience indicator system, the resilience indicator includes an ecological resilience indicator, which includes an ecological footprint indicator and an ecological risk indicator. The evaluation is performed by combining a fuzzy comprehensive evaluation method with a risk matrix method. Step S4 also includes the step of dynamic early warning of ecological resilience indicators: When the ecological resilience index is lower than the set warning threshold, a warning signal is issued.

5. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: In S5: solution evaluation and optimization step, a multi-objective evolutionary algorithm is also included: The non-dominated sorting genetic algorithm II is used to iteratively search and obtain the Pareto optimal solution set. The preference information entropy method is used to screen and sort the Pareto optimal solution set, and multiple candidate decision plans are obtained. At the same time, combined with the spatial analysis function of the geographic information system, the candidate decision plans are spatially visualized and analyzed.

6. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: In S6: Planning scheme formulation step, a dynamic land use zoning strategy based on scenario perception is generated. The scenario perception technology of artificial intelligence is used to perceive the development scenario and environmental changes of the industrial park in real time. According to different scenarios, the industrial park is divided into different dynamic zones. Different land use functions and development strategies are generated for each zone, and corresponding differentiated land use policies and management measures are generated.

7. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: In the step S1: Data collection and preprocessing, edge computing technology is used to locally process and analyze the collected data. Edge computing equipment is deployed at the data collection site to perform preliminary processing of the data in real time. At the same time, a federated learning algorithm is used to collaboratively analyze multi-source data to achieve knowledge sharing and model training between different data sources.

8. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: In S2: scenario setting step, there is also a policy model simulation step: Based on the method of combining system dynamics and machine learning, the development trend of industrial parks under different policy scenarios is simulated. The policy test bed method is used to test and evaluate different policy combinations, analyze the implementation effect and potential impact of the policy, and obtain policy simulation results. The policy simulation results are used as input conditions to generate land use scenarios.

9. The multi-scenario simulation-driven low-carbon industrial park land use flexibility planning method according to claim 1 is characterized in that: After step S6, there is also step S7: real-time evaluation step: During the implementation of the planning scheme, a performance evaluation and incentive mechanism based on blockchain will be established. The distributed ledger technology of blockchain will be used to record the implementation process and performance data of the planning scheme. The balanced scorecard method will be used to build a performance evaluation indicator system. The implementation effect of the planning scheme will be evaluated from four dimensions: finance, customers, internal processes, and learning and growth. Based on the evaluation results, incentive rewards will be automatically issued through smart contracts. At the same time, big data analysis technology will be used to conduct in-depth mining and analysis of performance evaluation data to provide decision support for the adjustment and optimization of the planning scheme. After step S6, it also includes real-time monitoring and dynamic adjustment steps. An intelligent monitoring network is deployed in the industrial park. The monitoring equipment realizes real-time transmission and sharing of data through 5G communication technology. Blockchain technology is used to build a data trusted storage and sharing platform. The real-time monitoring data is input into the online simulation model. When the deviation between the actual situation and the expected scenario exceeds the set threshold, a method based on rule reasoning and case reasoning is used to make dynamic adjustments. Rule reasoning makes decisions based on a preset rule library, and case reasoning searches for similar cases from the historical case library for reference. Using transfer learning technology, models trained in other similar industrial parks are migrated to the current industrial park.

10. A multi-scenario simulation-driven low-carbon industrial park flexible planning system, characterized by: include: The data collection and preprocessing module receives basic industrial park data collected by IoT sensor networks, satellite remote sensing, and drone aerial surveys, uses deep learning anomaly data detection algorithms to clean the data, and then normalizes the data; The scenario setting module runs a conditional generative adversarial network, takes different factors as input, generates land use scenarios, determines key driving factors and their value ranges, and uses the combination of the analytic hierarchy process (AHP) and the entropy weight method (EWM) to determine the weights of each driving factor. The multi-scenario simulation module builds a land use simulation model based on the improved cellular automation-multi-agent system coupled with system dynamics, simulates the dynamic feedback of multiple systems in the industrial park, and inputs it into the carbon emission dynamic accounting model to obtain the total carbon emissions of the park; The elasticity indicator system construction module is used to establish a dynamic land use elasticity indicator system including structure, function, time and resilience. The scheme evaluation and optimization module uses a multi-attribute decision-making method based on evidence theory to evaluate schemes, an improved particle swarm optimization algorithm to optimize schemes, and analyzes the sensitivity and robustness of schemes; The planning scheme formulation module adopts a combination of dynamic and rolling planning methods to divide the planning scheme into short-, medium- and long-term goals, clarify the spatial layout and functional positioning of various types of land, and generate ecological compensation mechanisms and low-carbon incentive policies; The feedback data collection module uses the data collection and preprocessing module to continue to receive implementation process and performance data during the execution of the planning scheme, and adopts the balanced scorecard method to build a performance evaluation indicator system to evaluate the implementation effect of the planning scheme from four dimensions: finance, customers, internal processes, and learning and growth.

Images