A land change simulation method and system considering multiple policy interactions and multi-level intelligent agents

By constructing a multi-level intelligent agent model that considers the interactions of the central government, urban construction, village collectives, farmland protection, and ecological protection, the problem of the lack of government-level consideration in existing models is solved, and more accurate land use simulation and urban expansion prediction are achieved.

CN118333475BActive Publication Date: 2026-11-17CHINA UNIV OF GEOSCIENCES (WUHAN)
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410534493.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-30
Publication Date
2026-11-17
Estimated Expiration
2044-04-30

AI Technical Summary

Technical Problem

Existing land use models lack consideration for different levels of government entities, making it impossible to accurately simulate the driving force and impact of government policies on land use. Furthermore, multi-entity models only consider a single land use type, failing to fully reflect the complexity and diversity of land use.

Method used

The system constructs four types of intelligent entities: central government, urban construction, village collectives, farmland protection, and ecological protection. It simulates land use change through multi-level interactions, taking into account the policy and behavioral preferences of governments at all levels, and uses geospatial data and decision-making models to simulate land conversion and use.

Benefits of technology

It improves the accuracy of land use simulation, enabling it to more accurately reflect the impact of multi-level government policies on land use, replicate the real land use change process, and enhance the accuracy of urban land expansion simulation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118333475B_ABST
    Figure CN118333475B_ABST
Patent Text Reader

Abstract

This invention discloses a land use change simulation method and system considering multiple policy interactions and multi-level intelligent agents. Relating to the field of land use change simulation, this method and system utilizes five types of intelligent agents across three levels to simulate the spatial allocation and adjustment of land use scale demand from macro to micro perspectives, and the spatial allocation mechanism of land use from temporal to spatial perspectives. By considering the behavioral preferences of different levels of government entities and the interactions of different land policies, it "replicates" the complex interactions and feedback mechanisms between government and its policies, the environment, and land use patterns, thereby constructing a simulation that more closely reflects the actual land use change process. Implementing the land use change simulation method and system considering multiple policy interactions provided by this invention can expand the research on urban land expansion mechanisms under multiple policy interactions and improve simulation accuracy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of land use change simulation, and more specifically, to a method and system for simulating land use change by a multi-level intelligent agent that considers multiple policy interactions. Background Technology

[0002] Land use / cover change is a key issue influencing global change and a core research area for international organizations such as the International Geosphere and Biosphere Programme and the Human Factors Programme on Global Change. Since the 21st century, my country's land use structure and spatial pattern have undergone profound changes. On the one hand, land has played a crucial role in my country's rapid economic growth, powerfully driving the rapid rise, development, and construction of cities. On the other hand, structural contradictions such as disordered and unbalanced land resource development, the coexistence of resource scarcity and inefficient use, and the spatial and temporal misallocation of regional resources have gradually emerged. Problems such as land pollution, environmental degradation, ecological damage, and biodiversity loss are becoming increasingly severe, becoming bottlenecks restricting sustainable economic and social development. Conducting research on the characteristics of land use / cover change, and exploring and analyzing the driving and guiding role of land policies and related planning on land use / cover change, is conducive to improving the accuracy and efficiency of regional land resource allocation and utilization, promoting sustainable land management and utilization, and is an important means to alleviate the current contradictions between land space development and protection, and to solve practical problems such as land pollution and land resource waste.

[0003] Land use process modeling can visualize the impact of various socio-economic processes on landscape patterns over a certain period and provide insights into the basic trends of urban development under different scenarios, serving as an important tool to support urban planning and policy-making. Currently, scholars both domestically and internationally have conducted extensive research and achieved fruitful results in modeling, with land use change simulation achieving a leap from quantitative simulation to spatiotemporal pattern simulation, and from single models to multiple coupled models. Cellular automata (CA) models and multi-agent / intelligent agent models (ABM), two "bottom-up" modeling methods, have become the mainstream models for current land use simulation due to their powerful spatiotemporal modeling capabilities. However, CA models focus on natural geographical environmental elements and cannot capture the impact of macro-social and human-driven factors on urban growth, nor can they accurately simulate human decision-making. Currently, most ABM modeling considers land use entities such as developers, enterprises, urban residents, and farmers, lacking consideration for land use management entities—government entities at different levels; furthermore, multi-agent land use models mostly consider only a single land use type, simulating land as construction land and non-construction land, lacking modeling for multiple land types such as arable land, forest land, and water bodies. As the main body of social public management, the government plays a leading role in constructing the basic institutional framework and related policies for land use and management through legal, economic, and administrative means such as taxation, financial systems, and fiscal measures. For example, the Land Administration Law, territorial spatial planning, the Basic Farmland Protection Regulations, and the balance between cultivated land occupation and compensation all influence the mode and degree of land use. A reasonable policy framework can promote sustainable land use practices or incentivize specific land use behaviors. Government behavior in the land use process is not undifferentiated. Due to the hierarchical nature of government agencies (central-provincial-municipal-county-township), the powers, scope, and development goals of governments at each level differ, leading to different land use objectives and intentions. This results in differentiated government behavior and differentiated land policy implementation, ultimately manifesting as different land use patterns and characteristics. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for simulating land change by considering the interaction of multiple policies and multi-level intelligent agents, which can expand the research on the mechanism of urban land expansion under the interaction of multiple policies and improve the simulation accuracy.

[0005] This invention provides a land change simulation method considering multiple policy interactions and a multi-level intelligent agent, comprising the following steps: S1: Constructing a central government intelligent agent; based on planning data, using the central government intelligent agent to obtain planning indicators and central government approval rate parameters; planning indicators include construction land planning indicators, cultivated land protection scale indicators, and ecological red line protection scale indicators; S2: Constructing an urban construction intelligent agent; using the urban construction intelligent agent to obtain expected urban development needs; S3: Constructing a village collective intelligent agent; using the village collective intelligent agent to obtain expected village development needs; S4: Based on the central government approval rate and expected urban development needs... S5: Based on the city's development needs and expected village development needs, allocate the annual construction land development scale to each village using the urban construction agent; S6: Based on the construction land development scale, conduct land acquisition activities using the urban construction agent and the village collective agent to obtain the new village construction land development scale; S7: Use the geospatial dataset to obtain the construction land development probability and initial land use status, and based on the village construction land development scale, use the village collective agent to propose land conversion requests based on the construction land development probability, generating the land conversion quantity; S8: Construct a farmland protection agent and generate the farmland conversion approval rate. Parameters; Based on the land conversion request, the farmland protection agent makes a behavioral decision on the land conversion request of the village collective agent. If the land conversion request is approved, the land use status is updated; if the land conversion request is rejected, return to step S6; S8: Construct an ecological protection agent and generate ecological land conversion approval rate parameters; Based on the land conversion request, the ecological protection agent makes a behavioral decision on the land conversion request of the village collective agent. If the land conversion request is approved, the land use status is updated; if the land conversion request is rejected, return to step S6; S9: Generate farmland using the farmland protection agent. The land reclamation rate parameter is used to obtain the amount of converted arable land and the amount of arable land reclaimed, and the land use status is updated; S10: Based on the amount of land converted and the land use status, the total amount of newly added construction land is obtained using the central government agent; Based on the total amount of newly added construction land and planning indicators, the new central government approval rate is obtained; S11: Steps S4-S10 are repeated for simulation. When the amount of land converted is not less than the scale of construction land development, or the developable land of the village collective agent is less than the preset value, or the simulated year exceeds the last year of the preset simulation, the model stops and a new land use status is obtained.

[0006] Furthermore, step S2 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions specifically includes: constructing an urban construction intelligent agent, and using the urban construction intelligent agent to obtain the expected urban development needs, as shown in the formula:

[0007] CityDemand = 0.5d Extrapolation +0.5d Markov

[0008] d Extrapolation =y(t),y(t)=a1t 2 +a2t+a

[0009]

[0010] Where CityDemand is the predicted amount of city-level construction land, d Extrapolation The amount of construction land is predicted based on trend extrapolation, d Markov It is based on the amount of construction land predicted by Markov, y(t) is a curve function of time t, a1 and a2 are the coefficients of the variables, and a is a constant term. and P represents the land use status at times t1 and t2. ij This represents the transition probability between land use types i and j.

[0011] Furthermore, step S4 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions specifically includes: allocating the annual construction land development scale for each village using an urban construction intelligent agent based on the central government approval rate, expected urban development needs, and expected village development needs, as shown in the formula:

[0012]

[0013]

[0014] In the formula, For the city construction intelligence agent, the land allocation coefficient for each village collective intelligence agent in a specified year t is given by CG. approval For approval rates, CityDemand t VillageDemand is a forecast of the city's expected development demand for a given year t. t,i For a given year t, the predicted development needs of the i-th village collective agent are given by ∑VillageDemand. t,i Allocation is the sum of the expected needs of all villages. t,i The development scale of construction land ultimately allocated to each village collective intelligent entity by the urban construction intelligent entity; the approval rate value is not less than 0 and not greater than 1.

[0015] Furthermore, step S5 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions specifically includes: based on the scale of construction land development, using urban construction agents and village collective agents to conduct land acquisition activities, obtaining the result of the village collective agreeing or refusing to sell the land; based on the scale of construction land development and the result of the village collective agreeing or refusing to sell the land, obtaining the new scale of village construction land development, as shown in the formula:

[0016]

[0017] In this context, SaleLand? represents the outcome of the village collective's decision to agree or disagree with the land sale; approve indicates approval, reject indicates rejection, and sell... reject Let r be the probability that the village collective intelligent agent rejects the land acquisition behavior of the urban construction intelligent agent. Let r be a random number between [0-1]. If the generated random number r is greater than sale reject If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

[0018] Furthermore, step S6 of the above-mentioned multi-level agent land change simulation method considering multiple policy interactions specifically includes: obtaining the probability of construction land development and the initial land use status using a geospatial dataset; based on the scale of village construction land development, using the village collective agent, converting the land allocated by the urban construction agent of the village collective agent in the corresponding year to the target quantity of construction land into construction land based on the construction land development probability, and obtaining the land conversion quantity and land conversion request, as shown in the formula:

[0019] DevP final,ij =α×DevP inital,ij +β×DevP local,ij +γ×P random,ij

[0020] Among them, DevP final,ij Let DevP be the final development probability of the construction land for pixel ij. inital,ij Let DevP be the initial probability of land development for pixel ij. local,ij Let P be the local development probability of pixel ij. random,ij α is a random number representing the uncertainty of land development, β is the weight of the initial development probability, γ is the weight of the local development probability, and γ is the weight of the random probability.

[0021] This invention also provides a system comprising the following modules: a planning indicator module, configured to: construct a central government intelligent agent, and obtain planning indicators and central government approval rate parameters based on planning data and utilizing the central government intelligent agent; the planning indicators include construction land planning indicators, cultivated land protection scale indicators, and ecological red line protection scale indicators; an expected urban development demand generation module, configured to: construct an urban construction intelligent agent, and obtain expected urban development demands using the urban construction intelligent agent; an expected village development demand generation module, configured to: construct a village collective intelligent agent, and obtain expected village development demands using the village collective intelligent agent; and a construction land development scale generation module, configured to: generate the scale of construction land based on the central government approval rate, expected urban development demands, and expected village development demands. The system utilizes urban construction intelligence to allocate annual construction land development scale to each village. The construction land development scale update module is configured to: based on the construction land development scale, conduct land acquisition activities using both urban construction intelligence and village collective intelligence to obtain the new village construction land development scale; the land conversion request module is configured to: obtain the construction land development probability and initial land use status using geospatial datasets, and based on the village construction land development scale, use the village collective intelligence to propose land conversion requests based on the construction land development probability, generating the land conversion quantity; the farmland protection decision-making module is configured to: construct a farmland protection intelligence and generate farmland conversion approval rate parameters; based on the land conversion requests, use the farmland protection intelligence to... The collective intelligence agent makes behavioral decisions regarding land conversion requests. If the request is approved, the land use status is updated; if rejected, the process returns to the land conversion request module. The ecological protection decision module is configured to: construct an ecological protection intelligence agent and generate ecological land conversion approval rate parameters; based on the land conversion request, utilize the ecological protection intelligence agent to make behavioral decisions regarding the village collective intelligence agent's land conversion request. If the request is approved, the land use status is updated; if rejected, the process returns to the land conversion request module. The cultivated land reclamation quantity module is configured to: utilize the cultivated land protection intelligence agent to generate cultivated land reclamation rate parameters, obtain the quantity of converted cultivated land and the quantity of cultivated land reclaimed, and update the land use status. The central government approval rate... The update module is configured to: obtain the cumulative total amount of newly added construction land based on the land conversion quantity and land use status using the central government agent; and obtain the new central government approval rate based on the cumulative total amount of newly added construction land and planning indicators. The iteration module is configured to: simulate the process of the repeated construction land development scale generation module, construction land development scale update module, land conversion request module, farmland protection decision module, ecological protection decision module, farmland reclamation quantity module, and central government approval rate update module. When the land conversion quantity is not less than the construction land development scale, or the developable land of the village collective agent is less than the preset value, or the simulated year exceeds the last year of the preset simulation, the model stops and a new land use status is obtained.

[0022] Furthermore, the specific configuration of the expected urban development demand generation module of the above system is as follows: Construct an intelligent agent for urban construction, and use this agent to obtain the expected urban development demand, as shown in the formula:

[0023] CityDemand = 0.5d Extrapolation +0.5d Markov

[0024] d Extrapolation =y(t),y(t)=a1t 2 +a2t+a

[0025]

[0026] Where CityDemand is the predicted amount of city-level construction land, d Extrapolation The amount of construction land is predicted based on trend extrapolation, d Markov It is based on the amount of construction land predicted by Markov, y(t) is a curve function of time t, a1 and a2 are the coefficients of the variables, a is the constant term, and S t1 and S t2 P represents the land use status at times t1 and t2. ij This represents the transition probability between land use types i and j.

[0027] Furthermore, the specific configuration of the construction land development scale generation module of the above system is as follows: based on the central government's approval rate, expected urban development needs, and expected village development needs, the urban construction agent allocates the annual construction land development scale to each village, as shown in the formula:

[0028]

[0029] Allocation t,i =f allocationt ×VillageDemand t,i

[0030] In the formula, f allocationt For the city construction intelligence agent, the land allocation coefficient for each village collective intelligence agent in a specified year t is given by CG. approval For approval rates, CityDemand t VillageDemand is a forecast of the city's expected development demand for a given year t. t,i For a given year t, the predicted development needs of the i-th village collective agent are given by ∑VillageDemand. t,i Allocation is the sum of the expected needs of all villages. t,iThe development scale of construction land ultimately allocated to each village collective intelligent entity by the urban construction intelligent entity; the approval rate value is not less than 0 and not greater than 1.

[0031] Furthermore, the specific configuration of the construction land development scale update module of the above system is as follows: Based on the construction land development scale, land acquisition activities are carried out using urban construction agents and village collective agents to obtain the result of the village collective agreeing or refusing to sell the land. Based on the construction land development scale and the result of the village collective agreeing or refusing to sell the land, a new village construction land development scale is obtained, as shown in the formula:

[0032]

[0033] In this context, SaleLand? represents the outcome of the village collective's decision to agree or disagree with the land sale; approve indicates approval, reject indicates rejection, and sell... reject Let r be the probability that the village collective intelligent agent rejects the land acquisition behavior of the urban construction intelligent agent. Let r be a random number between [0-1]. If the generated random number r is greater than sale reject If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

[0034] Furthermore, the specific configuration of the land conversion request module of the above system is as follows: Using a geospatial dataset, the development probability of construction land and the initial land use status are obtained. Based on the village's construction land development scale, and utilizing the village collective intelligence agent, the land allocated by the urban construction intelligence agent of the corresponding year to the village collective intelligence agent for the same target quantity is converted into construction land based on the construction land development probability, thus obtaining the land conversion quantity and land conversion request, as shown in the formula:

[0035] DevP final,ij =α×DevP inital,ij +β×DevP local,ij +γ×P random,ij

[0036] Among them, DevP final,ij Let DevP be the final development probability of the construction land for pixel ij. inital,ij Let DevP be the initial probability of land development for pixel ij. local,ij Let P be the local development probability of pixel ij. random,ij α is a random number representing the uncertainty of land development, β is the weight of the initial development probability, γ is the weight of the local development probability, and γ is the weight of the random probability.

[0037] The land change simulation method and system considering multiple policy interactions and multi-level intelligent agents provided by this invention have the following beneficial effects:

[0038] Based on the hierarchical structure of my country's planning system and the graded management model of government land use, this invention uses five types of intelligent agents across three levels to simulate the spatial allocation mechanism of land use from macro to micro levels, and from time to space, thereby simulating the spatiotemporal changes in regional land use patterns. By considering the behavioral preferences of different levels of government entities and the interaction of different land policies, it "replicates" the complex interactions and feedback mechanisms between government and its policies, the environment, and land use patterns. This constructs a model that more closely reflects the actual land use change process, helping to quantitatively and visually express the interaction between multi-level governments and policies behind urban land change (urban expansion). This is beneficial for subsequent subject modeling research based on policy interaction and the study of the mechanism of human-land system coupling. It can better replicate and "reproduce" the spatial pattern of urban land use change, improving the accuracy of urban land expansion simulation. Attached Figure Description

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0040] Figure 1 This is a flowchart of the land change simulation method for multi-level intelligent agents that considers multiple policy interactions provided by the present invention;

[0041] Figure 2 This is a schematic diagram of the land change simulation system module and its components that consider multiple policy interactions and multi-level intelligent agents provided by the present invention.

[0042] Figure 3 This is a class diagram of a multi-level intelligent agent program design provided by the present invention;

[0043] Figure 4 This invention provides a schematic diagram illustrating the interaction between multiple agents and the environment in the land use decision-making process.

[0044] Figure 5 This is a schematic diagram of the land change simulation technology process provided by the present invention, which considers multiple policy interactions and multi-level intelligent agents.

[0045] Figure 6 This is a schematic diagram of the initial land use unit and its candidate development unit provided by the present invention;

[0046] Figure 7 This is a schematic diagram illustrating the changes in land use unit types and their decision support intentions provided by the present invention. Detailed Implementation

[0047] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0048] Figure 1 A schematic diagram of the land change simulation method considering multiple policy interactions and multi-level agent interactions, as shown in this embodiment, is illustrated. In this embodiment, the land change simulation method considering multiple policy interactions and multi-level agent interactions includes the following steps:

[0049] S1: Construct a central government intelligent agent. Based on the planning data, use the central government intelligent agent to obtain planning indicators and central government approval rate parameters. The planning indicators include construction land planning indicators, cultivated land protection scale indicators and ecological red line protection scale indicators.

[0050] S2: Construct an intelligent agent for urban construction and use it to obtain the expected urban development needs;

[0051] Specifically, step S2 of the above-mentioned land change simulation method considering multiple policy interactions and multi-level intelligent agents includes: constructing an urban construction intelligent agent, and using the urban construction intelligent agent to obtain the expected urban development needs, as shown in the formula:

[0052] CityDemand = 0.5d Extrapolation +0.5d Markov

[0053] d Extrapolation =y(t),y(t)=a1t 2 +a2t+a

[0054]

[0055] Where CityDemand is the predicted amount of city-level construction land, d Extrapolation The amount of construction land is predicted based on trend extrapolation, d Markov It is based on the amount of construction land predicted by Markov, y(t) is a curve function of time t, a1 and a2 are the coefficients of the variables, and a is a constant term. and P represents the land use status at times t1 and t2. ij This represents the transition probability between land use types i and j;

[0056] S3: Construct a village collective intelligent agent and utilize it to obtain the expected village development needs;

[0057] Specifically, step S3 of the above-mentioned land change simulation method considering multiple policy interactions and multi-level intelligent agents includes: constructing a village collective intelligent agent, and using the village collective intelligent agent to obtain the expected village development needs, as shown in the formula:

[0058] VillageDemand i =y(t),y(t)=b 1i t 2 +b 2i t+b i

[0059] Where VillageDemand represents the projected construction land demand for each village, y(t) is a curve function of time t, and b 1i and b 2i b represents the coefficient of the variable corresponding to village i. i It is a constant term;

[0060] S4: Based on the central government's approval rate, expected urban development needs, and expected village development needs, the urban construction intelligence agent allocates the annual construction land development scale to each village.

[0061] Specifically, step S4 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions includes: allocating the annual construction land development scale for each village using an urban construction intelligent agent based on the central government approval rate, expected urban development needs, and expected village development needs, as shown in the formula:

[0062]

[0063]

[0064] In the formula, For the city construction intelligence agent, the land allocation coefficient for each village collective intelligence agent in a specified year t is given by CG. approval For approval rates, CityDemand t VillageDemand is a forecast of the city's expected development demand for a given year t. t,i For a given year t, the predicted development needs of the i-th village collective agent are given by ∑VillageDemand. t,i Allocation is the sum of the expected needs of all villages. t,i The final scale of construction land development allocated to each village collective intelligence by the urban construction intelligence agent; the approval rate value is not less than 0 and not greater than 1.

[0065] S5: Based on the scale of construction land development, land acquisition activities are carried out using urban construction intelligence and village collective intelligence to obtain the new scale of village construction land development.

[0066] Specifically, step S5 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions includes: based on the scale of construction land development, conducting land acquisition activities using urban construction agents and village collective agents, obtaining the result of the village collective agreeing or refusing to sell the land, and obtaining the new village construction land development scale based on the scale of construction land development and the result of the village collective agreeing or refusing to sell the land, as shown in the formula:

[0067]

[0068] In this context, SaleLand? represents the outcome of the village collective's decision to agree or disagree with the land sale; approve indicates approval, reject indicates rejection, and sell... reject Let r be the probability that the village collective intelligent agent rejects the land acquisition behavior of the urban construction intelligent agent. Let r be a random number between [0-1]. If the generated random number r is greater than sale reject If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

[0069] S6: Use geospatial datasets to obtain the probability of construction land development and the initial land use status. Based on the scale of village construction land development, use village collective intelligence to propose land conversion requests based on the probability of construction land development and generate the number of land conversions.

[0070] Specifically, step S6 of the above-mentioned multi-level agent land change simulation method considering multiple policy interactions includes: obtaining the probability of construction land development and the initial land use status using a geospatial dataset; based on the scale of village construction land development, using the village collective agent, converting the land allocated by the urban construction agent of the village collective agent in the corresponding year to the village collective agent into construction land based on the probability of construction land development, and obtaining the land conversion quantity and land conversion request, as shown in the formula:

[0071] DevP final,ij =α×DevP inital,ij +β×DevP local,ij +γ×P random,ij

[0072] Among them, DevP final,ij Let DevP be the final development probability of the construction land for pixel ij. inital,ij Let DevP be the initial probability of land development for pixel ij. local,ij Let P be the local development probability of pixel ij. random,ijα is a random number representing the uncertainty of land development, β is the weight of the initial development probability, γ is the weight of the local development probability, and γ is the weight of the random probability.

[0073] S7: Construct a farmland protection intelligent agent and generate farmland conversion approval rate parameters; based on the land conversion request, use the farmland protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent. If the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to step S6.

[0074] Specifically, step S7 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions includes: constructing a farmland protection intelligent agent and generating farmland conversion approval rate parameters; based on land conversion requests, using the farmland protection intelligent agent to make behavioral decisions regarding the land conversion requests of the village collective intelligent agent; if the land conversion request is approved, the land use status is updated; if the land conversion request is rejected, the process returns to step S6 as shown in the formula:

[0075]

[0076] Where Cconversion? represents the result of the farmland protection agent's agreement or rejection of land conversion, with approve indicating approval and reject indicating rejection. FPC approval To increase the approval rate of intelligent agents for farmland protection;

[0077] S8: Construct an ecological protection intelligent agent and generate ecological land conversion approval rate parameters; based on the land conversion request, use the ecological protection intelligent agent to make behavioral decisions on the land conversion request of the village collective intelligent agent. If the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to step S6.

[0078] Specifically, step S8 of the above-mentioned multi-level intelligent agent land change simulation method considering multiple policy interactions includes: constructing an ecological protection intelligent agent and generating ecological land conversion approval rate parameters; based on land conversion requests, using the ecological protection intelligent agent to make behavioral decisions regarding the land conversion requests of the village collective intelligent agent; if the land conversion request is approved, the land use status is updated; if the land conversion request is rejected, the process returns to step S6, as shown in the formula:

[0079]

[0080] Among them, Econversion? represents the outcome of the ecological protection agent's agreement or rejection of land conversion, with approve indicating approval and reject indicating rejection. EPC approval For the approval rate of ecological protection intelligent agents;

[0081] S9: Use the farmland protection intelligent agent to generate farmland reclamation rate parameters, obtain the amount of farmland converted and the amount of farmland reclaimed, and update the land use status;

[0082] Specifically, step S9 of the above-mentioned land change simulation method considering multiple policy interactions and multi-level intelligent agents includes: generating farmland reclamation rate parameters using a farmland protection intelligent agent, obtaining the amount of farmland to be converted and the amount of farmland reclaimed, and updating the land use status, as shown in the formula:

[0083]

[0084] in, FPC_reclaim represents the amount of arable land reclaimed in the specified year t; FPC_reclaim represents the arable land reclamation rate. This represents the amount of arable land occupied by construction in the study area for the corresponding year.

[0085] S10: Based on the amount of land conversion and land use status, the cumulative total amount of newly added construction land is obtained using the central government's intelligent agent; based on the cumulative total amount of newly added construction land and planning indicators, the new central government approval rate is obtained.

[0086] Specifically, step S10 of the above-mentioned land change simulation method considering multiple policy interactions and multi-level intelligent agents includes: obtaining the cumulative total amount of newly added construction land using the central government intelligent agent based on the amount of land conversion and land use status; and obtaining the new central government approval rate based on the cumulative total amount of newly added construction land and planning indicators, as shown in the formula:

[0087]

[0088] Among them, CG approval Represents the central government's approval rate, count construction It is the cumulative amount of land converted into construction land starting from the initial year, quota construction This represents the planning indicators set by the central government's intelligent agent at the initial planning stage. The model stops if the central government's approval rate is 0.

[0089] S11: Repeat steps S4-S10 to perform the simulation. When the number of land conversions is not less than the scale of construction land development, or the developable land of the village collective intelligence is less than the preset value, or the simulation year exceeds the last year of the preset simulation, the model stops and a new land use status is obtained.

[0090] Figure 2 The modular composition and contents of the system of the present invention are illustrated. Specifically, the present invention also provides a system comprising the following modules:

[0091] The planning indicators module is configured as follows: Construct a central government intelligent agent, and based on the planning data, use the central government intelligent agent to obtain planning indicators and central government approval rate parameters. The planning indicators include construction land planning indicators, cultivated land protection scale indicators and ecological red line protection scale indicators.

[0092] The module for generating anticipated urban development needs is configured as follows: construct an intelligent agent for urban construction, and use the intelligent agent for urban construction to obtain anticipated urban development needs;

[0093] The module for generating expected village development needs is configured as follows: construct a village collective intelligent agent, and use the village collective intelligent agent to obtain the expected village development needs;

[0094] The construction land development scale generation module is configured as follows: based on the central government's approval rate, expected urban development needs, and expected village development needs, the urban construction intelligence agent allocates the annual construction land development scale to each village.

[0095] The module for updating the scale of construction land development is configured as follows: Based on the scale of construction land development, land acquisition activities are carried out using urban construction intelligence agents and village collective intelligence agents to obtain the new scale of village construction land development.

[0096] The land conversion request module is configured to: obtain the development probability of construction land and the initial land use status using a geospatial dataset; based on the development scale of village construction land, use the village collective intelligent agent to propose a land conversion request based on the development probability of construction land and generate the number of land conversions;

[0097] The farmland protection decision-making module is configured as follows: construct a farmland protection intelligent agent and generate farmland conversion approval rate parameters; based on the land conversion request, use the farmland protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent; if the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to the land conversion request module.

[0098] The ecological protection decision-making module is configured as follows: construct an ecological protection intelligent agent and generate ecological land conversion approval rate parameters; based on the land conversion request, use the ecological protection intelligent agent to make behavioral decisions on the land conversion request of the village collective intelligent agent; if the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to the land conversion request module.

[0099] The cultivated land reclamation quantity module is configured as follows: using the cultivated land protection intelligent agent to generate cultivated land reclamation rate parameters, obtain the converted cultivated land quantity and cultivated land reclamation quantity, and update the land use status;

[0100] The central government approval rate update module is configured as follows: based on the number of land conversions and land use status, the cumulative total amount of newly added construction land is obtained using the central government intelligent agent; based on the cumulative total amount of newly added construction land and planning indicators, the new central government approval rate is obtained.

[0101] The iterative module is configured as follows: the simulation process of the repeated construction land development scale generation module, construction land development scale update module, land conversion request module, farmland protection decision module, ecological protection decision module, farmland reclamation quantity module and central government approval rate update module. When the land conversion quantity is not less than the construction land development scale, or the developable land of the village collective intelligence is less than the preset value, or the simulated year exceeds the last year of the preset simulation, the model stops and a new land use state is obtained.

[0102] Specifically, the configuration of the above-mentioned system's expected urban development demand generation module is as follows: Construct an intelligent agent for urban construction; utilize this intelligent agent to obtain the expected urban development demand, as shown in the formula:

[0103] CityDemand = 0.5d Extrapolation +0.5d Markov

[0104] d Extrapolation =y(t),y(t)=a1t 2 +a2t+a

[0105]

[0106] Where CityDemand is the predicted amount of city-level construction land, d Extrapolation The amount of construction land is predicted based on trend extrapolation, d Markov It is based on the amount of construction land predicted by Markov, y(t) is a curve function of time t, a1 and a2 are the coefficients of the variables, and a is a constant term. and P represents the land use status at times t1 and t2. ij This represents the transition probability between land use types i and j.

[0107] Specifically, the configuration of the construction land development scale generation module of the above system is as follows: based on the central government approval rate, expected urban development needs, and expected village development needs, the urban construction agent allocates the annual construction land development scale to each village, as shown in the formula:

[0108]

[0109]

[0110] In the formula, For the city construction intelligence agent, the land allocation coefficient for each village collective intelligence agent in a specified year t is given by CG. approval For approval rates, CityDemand t VillageDemand is a forecast of the city's expected development demand for a given year t. t,i For a given year t, the predicted development needs of the i-th village collective agent are given by ∑VillageDemand. t,i Allocation is the sum of the expected needs of all villages. t,i The development scale of construction land ultimately allocated to each village collective intelligent entity by the urban construction intelligent entity; the approval rate value is not less than 0 and not greater than 1.

[0111] Specifically, the configuration of the above-mentioned system's construction land development scale update module is as follows: Based on the construction land development scale, land acquisition activities are conducted using urban construction agents and village collective agents to obtain the result of the village collective agreeing or refusing to sell the land. Based on the construction land development scale and the result of the village collective agreeing or refusing to sell the land, a new village construction land development scale is obtained, as shown in the formula:

[0112]

[0113] In this context, SaleLand? represents the outcome of the village collective's decision to agree or disagree with the land sale; approve indicates approval, reject indicates rejection, and sell... reject Let r be the probability that the village collective intelligent agent rejects the land acquisition behavior of the urban construction intelligent agent. Let r be a random number between [0-1]. If the generated random number r is greater than sale reject If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

[0114] Specifically, the land conversion request module of the above system is configured as follows: Using a geospatial dataset, the development probability of construction land and the initial land use status are obtained. Based on the village's construction land development scale, and utilizing the village collective agent, the land allocated by the urban construction agent of the corresponding year to the village collective agent for the same target quantity is converted into construction land based on the construction land development probability. This yields the land conversion quantity and the land conversion request, as shown in the formula:

[0115] DevP final,ij =α×DevP inital,ij +β×DevP local,ij +γ×P random,ij

[0116] Among them, DevP final,ij Let DevP be the final development probability of the construction land for pixel ij. inital,ij Let DevP be the initial probability of land development for pixel ij. local,ij Let P be the local development probability of pixel ij. random,ij α is a random number representing the uncertainty of land development, β is the weight of the initial development probability, γ is the weight of the local development probability, and γ is the weight of the random probability.

[0117] In some embodiments, the above-described method for simulating land change by multi-level agents considering multiple policy interactions can be implemented in the following ways.

[0118] The method in this embodiment employs the following technical solution: a multi-level agent land change simulation model that considers the interaction of multiple policies, specifically including the following steps:

[0119] Step 1: Create five types of intelligent agents across three levels. The first level consists of the central government (CG) intelligent agent, the second level consists of the local government intelligent agent, and the third level consists of the village collective intelligent agent (VA). The local government is further subdivided into three different intelligent agents based on the different functions of each department: urban development commission (UDC), farmland protect commission (FPC), and ecological protect commission (EPC).

[0120] Figure 3The multi-level intelligent agent programming class diagram created by this invention is shown. Specifically, it is based on object-oriented programming (OOP) and abstracts each type of intelligent agent into a class. The characteristics and behaviors of the intelligent agents are abstracted into the attributes and methods of the class, respectively. In the specific simulation process, the program will create instances of the class based on information from the real world. The instances will realize the simulated interaction between intelligent agents through the message passing mechanism. Among them, (1) CG intelligent agent class: its goal is to formulate land space planning by comprehensively considering multiple dimensions such as national food security, ecological security, and social development, and promote the sustainable development of land space. The specific behavior in this model is to be responsible for formulating planning indicators and to regulate the overall development scale of urban construction land based on the CG behavior decision model. It has attribute variables such as construction land planning indicators, cultivated land protection scale indicators, ecological red line protection scale indicators, and CG approval rate. (2) UDC intelligent agent class: its goal is to promote the rapid development of cities and economic construction. The specific behavior in this model is to predict urban land development needs, carry out urban land quantity allocation and land expropriation activities, and has variable attributes such as expected urban development needs, construction land development scale, and allocation coefficient. (3) VA agent class: Its goal is to maximize the benefits of village collectives based on individual behavior. The specific behavior in this model is to predict village land development needs, respond to UDC land expropriation activities, make decisions based on the land expropriation decision model, and dominate the spatial allocation of land, that is, to issue land conversion requests to FPC and EPC: the land conversion process is the process of converting non-construction land into construction land. It has variable attributes such as expected village development needs, construction land allocation target scale, and land expropriation rejection rate. (4) FPC agent class: Its goal is to protect permanent basic farmland and balance the occupation and replenishment of cultivated land. The specific behavior in this model is to make a support decision on whether to agree to the land conversion based on the VA's land conversion request, and at the same time dominate the cultivated land reclamation activities based on the amount of cultivated land converted into construction land. It has attribute variables such as FPC approval rate, amount of cultivated land reduction, and cultivated land reclamation rate. (5) EPC Agent Class: Its goal is to protect ecologically significant land. In this model, its specific behavior is to make a decision on whether to approve land conversion based on the VA's land conversion request. It has attribute variables such as EPC approval rate.

[0121] Step 2: At the beginning of a planning period T (t∈T), the central government agent (CG) sets binding targets for the scale of permanent basic farmland protection, the scale of newly added construction land and the scale of ecological red line protection up to the target year, based on relevant plans or related planning data issued by the state, provinces and cities.

[0122] Step 3: The Urban Construction Intelligence Agent (UDC) predicts the expected annual urban development demand, i.e., the expected amount of new construction land in the city. The formula is as follows:

[0123] CityDemand = 0.5d Extrapolation +0.5d Markov #(1)

[0124] d Extrapolation =y(t),y=a1t 2 +a2t+a#(2)

[0125]

[0126] Step 4: The Village Collective Intelligence Agent (VA) predicts the expected development needs of each village each year (the amount of newly added construction land for each village). The formula is as follows:

[0127] VillageDemand i =y(t),y=b 1i t 2 +b 2i t+b i #(4)

[0128] At each time step:

[0129] Step 5: The Urban Construction Agent (UDC) assigns the annual village development target, i.e., the target scale of construction land, to each village (VA) according to the land quantity allocation sub-model, as shown below:

[0130]

[0131]

[0132] Step 6: Based on the results of Step 5, UDC and VA conduct land acquisition activities, making a decision through a land acquisition sub-model (based on the parameter land acquisition rejection rate): If the land acquisition sub-model results in a rejection of the land sale, VA will reject UDC's land acquisition, and the village's construction land target for this year will be set to 0. The land acquisition sub-model is represented as follows:

[0133]

[0134] Step 7: If the target size of construction land allocated to the VA by UDC is greater than 0, the village collective agent (VA) selects land equivalent to the target amount allocated to the VA by UDC in the corresponding year and converts it into construction land according to the land spatial allocation sub-model. The land spatial allocation sub-model is represented as follows:

[0135] DevP final,ij =α×DevPinital,ij +β×DevP local,ij +γ×P random,ij #(9)

[0136] Step 8: The Farmland Protection (FPC) agent makes a behavioral decision regarding the land conversion request from the Village Collective (VA) agent, either agreeing to or rejecting the land conversion process (i.e., the process of converting farmland pixels into construction land pixels). This process is regulated by a 0-1 parameter, FPC-approval. If the decision sub-model results in agreement, the land use status is updated; otherwise, the process returns to Step 7. The FPC decision behavior is represented as follows:

[0137]

[0138] Step 9: The Ecological Protection (EPC) agent makes a behavioral decision regarding the land conversion request from the Village Collective (VA) agent, either agreeing to or rejecting the land conversion process (i.e., the process of converting ecological land pixels such as forest land, grassland, and water areas into construction land pixels). This process is regulated by a 0-1 parameter, EPC-approval. If the decision sub-model results in agreement, the land use status is updated; otherwise, the process returns to Step 7. The EPC decision behavior is represented as follows:

[0139]

[0140] Step 10: The Farmland Conservation (FPC) agent performs farmland replenishment and reclamation based on the farmland reclamation sub-model. The amount of farmland replenished is determined by the parameter FPC-reclaim. The farmland reclamation sub-model is represented as follows:

[0141]

[0142] Step 11: Central Government Agent (CG) Behavioral Decision-Making Model: Calculate the total cumulative amount of newly added construction land in the city. If the total cumulative amount of newly added construction land exceeds the planning target set by the CG for the planning period, the CG will adjust the approval rate for next year (CG-approval). If the CG approval rate is 0, the model stops.

[0143]

[0144] Step 12: Repeat steps 5-11 until the number of land conversions meets the construction land target size given to the VA by the UDC, or there is not enough developable land in the administrative village, or the model stops running in the last year T of the preset simulation, and the land use status is updated.

[0145] In some embodiments, the above-described method for simulating land change by multi-level agents considering multiple policy interactions can be implemented in the following ways.

[0146] Based on the hierarchical structure of my country's planning system and the graded management model of government land use, this embodiment provides a land change simulation method (Multilevel Agent-based model, MLABM) that couples multiple land policies and multi-level agent system interactions. By combining top-down agent design with bottom-up operational mechanisms, it replicates and reproduces the complex interactions and feedback mechanisms between government entities, the environment, and land use spatial entities at different levels. (See...) Figure 4 .

[0147] For detailed steps of this embodiment, see [link to example]. Figure 5 The initial settings are as follows:

[0148] Step 1: Create five types of intelligent agents across three levels. The first level consists of the central government (CG) intelligent agent, the second level consists of the local government intelligent agent, and the third level consists of the village collective intelligent agent (VA). The local government is further subdivided into three different intelligent agents based on the different functions of each department: urban development commission (UDC), farmland protect commission (FPC), and ecological protect commission (EPC).

[0149] Step 2: The central government intelligence agency (CG) sets binding targets for the scale of permanent basic farmland protection, the scale of newly added construction land, and the scale of ecological red line protection by the target year (a planning cycle, such as 2020-2035).

[0150] Step 3: The Urban Construction Agent (UDC) predicts the expected annual urban development demand, i.e., the expected amount of new construction land in the city. Commonly used methods for predicting urban construction land development demand include population-construction land standard quota prediction, economic mathematical models based on trend extrapolation, linear prediction, regression models, Markov prediction, and neural network algorithm prediction. This example uses trend extrapolation and Markov prediction. The formula is as follows:

[0151] CityDemand = 0.5d Extrapolation +0.5d Markov #(1)

[0152] d Extrapolation =y(t),y=a1t 2 +a2t+a#(2)

[0153]

[0154] In the formula, CityDemand is the predicted amount of city-level construction land. d Extrapolation The amount of construction land is predicted based on trend extrapolation, d Markov This is based on the amount of land for construction predicted by Markov. Extrapolation It is a curve function with respect to time t, where a1 and a2 represent the coefficients of the variable, and a is a constant term. and This represents the land use status at times t1 and t2. (P) ij This represents the probability of transition between land use types.

[0155] Step 4: The Village Collective Agent (VA) predicts the expected development needs of each village each year (the amount of newly added construction land for each village). In this example, trend extrapolation is used to predict the expected construction land needs of each Village Collective Agent, using the following formula:

[0156] VillageDemand i =y(t),y=b 1i t 2 +b 2i t+b i #(4)

[0157] In the formula, VillageDemand represents the projected construction land demand for each village. VillageDemand is a curve function of time t, and b 1i and b 2i b represents the coefficient of the variable corresponding to village i. i It is a constant term.

[0158] Each time step (year):

[0159] Step 5: The Urban Construction Agent (UDC), based on the approval rate of the Central Government Agent (CG), the expected development needs of the city, and the expected development needs of the villages, allocates the annual village development target, i.e., the target scale of construction land, to each village (VA) according to the land quantity allocation sub-model. The land quantity allocation sub-model is represented as follows:

[0160]

[0161]

[0162] In the formula, This is the construction land allocation coefficient of UDC to VA in a specified year t. CG approval It's an approval rate between 0 and 1, used to adjust the scale of construction land use for the following year. CG approvalThe initial value is set to 1. When the total accumulated construction land exceeds the planning target, CG will adjust this value to be less than 1 (CG behavioral decision model, see below). CityDemand t VillageDemand represents the projected urban development demand for a specified year (Formula 1). t,i The projected development demand of the village for a specified year (Formula 4). ∑VillageDemand t,i It is the sum of the anticipated needs of all villages. (Allocation) t,i This represents the target size of construction land ultimately allocated to each village by the UDC.

[0163] Step 7: Each village collective agent (VA) makes a decision based on the land acquisition sub-model: whether to agree to UDC's land acquisition plan. Specifically, a random number between 0 and 1 is generated. If the random number is less than the VA's land acquisition agreement rate, the VA will reject UDC's land acquisition, and the village's allocation target for this year will be set to 0. The land acquisition sub-model is shown below:

[0164]

[0165] In this context, SaleLand? represents the outcome of VA's agreement / rejection to sell the land, and sale... reject This represents the probability that villagers will refuse UDC's land expropriation. r is a random number between 0 and 1. If the generated random number r is greater than 'sale', then... reject If the village agrees, it will agree to UDC's land expropriation; otherwise, it will refuse.

[0166] Step 8: If the village's allocation target is greater than 0, the village collective agent (VA) selects undeveloped pixels equivalent to the village's allocation target number for the corresponding year as undeveloped pixels / land units (pixels that may potentially be converted into construction land). See [link to relevant documentation]. Figure 6 The selection rule is based on sorting the probability values ​​of construction land development from largest to smallest, i.e., the land spatial allocation sub-model. Step 8's specific implementation includes the following sub-steps:

[0167] Step 8.1—Initial Development Probability Assessment: In this embodiment, the initial development probability of urban construction for each raster cell is comprehensively assessed based on natural geographical, socio-economic, and locational factors such as transportation. Commonly used natural geographical factors mainly include topography, such as elevation, slope, and NDVI (National Distance Index). Commonly used socio-economic factors include population, GDP, and point-of-interest density. Commonly used locational and transportation factors include: distance to city, county, and town centers; distance to national highways, provincial highways, expressways, ports, and other transportation facilities; and distance to public service facilities such as schools and hospitals. Commonly used assessment methods mainly include evaluation index methods, statistical learning models (support vector machines, least squares support vector machines), ensemble learning models (such as random forests, deep forests), and neural network models (such as multilayer perceptrons, artificial neural networks, deep learning, and convolutional neural networks).

[0168] Step 8.2—Local Development Probability (Neighborhood Effect) Assessment: In this embodiment, the neighborhood land use effect is considered, regarded as one of the important factors affecting the transformation of pixel status. The more neighborhood pixels around a given pixel, the greater the probability that it will be converted into construction land. The formula is expressed as:

[0169]

[0170] DevP local,ij ∑ represents the local development probability of a pixel. N×N The count is the number of construction land units in the neighborhood, where N is the number of neighborhood units, using a 3×3 neighborhood window.

[0171] Step 8.3—Evaluation of the Final Development Probability of Construction Land: The development probability of construction land is jointly determined by natural conditions, socio-economic factors, transportation location conditions, the neighborhood effect of the construction land, and random factors. The formula is expressed as follows:

[0172] DevP final,ij =α×DevP inital,ij +β×DevP local,ij +γ×P random,ij #(9)

[0173] In the formula, DevP final,ij DevP represents the final development probability of the construction land for pixel ij. inital,ij The initial probability of land development for a given element (ij) is determined by driving factors such as natural conditions, socio-economic factors, and transportation location conditions; DevP local,ij Let P be the local development probability of pixel ij, i.e., the neighborhood effect. The more construction land pixels surrounding pixel ij, the greater the probability that it will be converted into construction land, and vice versa; random,ijThis is a random number between 0 and 1, used to represent the uncertainty of land development. Here, α, β, and γ are the weights of the initial development probability, the local development probability, and the random probability, respectively, and α + β + γ = 1.

[0174] Step 9: The Farmland Conservation (FPC) agent responds to the raster cell selected by the Village Collective (VA) agent, agreeing or rejecting the land conversion process (the process of converting farmland cells to construction land cells), and updates the land use status. See [link to relevant documentation]. Figure 7 Step 9 includes the following sub-steps:

[0175] Step 9.1: If the cell selected by the VA for development is not a basic farmland (PPF), then the FPC supports the VA's land conversion.

[0176] Step 9.2: If the VA selects a basic farmland (PPF) for development, a random number between 0 and 1 is generated. If the random number is less than the FPC approval rate (FPC_approval), the VA's land conversion is rejected; otherwise, it is supported. This decision model is represented as follows:

[0177]

[0178] Where Cconversion? represents the result of the FPC agent agreeing to or rejecting the land conversion, and FPC... approval The approval rate of the FPC agent is calibrated using a multi-objective optimization algorithm.

[0179] Step 9.3: If the FPC rejects the VA's land conversion, the VA repeats steps 8 and 9 until the number of undeveloped cells meets the target construction land size allocated by the UDC to the VA.

[0180] Step 10: The Ecological Protection (EPC) agent responds to the raster cells selected by the Village Collective (VA) agent, agreeing or rejecting the land conversion process (the process of converting ecological land cells such as forest land, grassland, and water areas into construction land cells), and updates the land use status. See [link to relevant documentation]. Figure 7 Step 10 includes the following sub-steps:

[0181] Step 10.1: If the cell selected by VA for development is not an Ecological Protection Red Line (ERL) cell, then EPC supports the land conversion of VA.

[0182] Step 10.2: If the pixel selected by VA for development is an Ecological Protection Red Line (ERL) pixel, then a random number between 0 and 1 is generated. If the random number is less than the EPC approval rate (EPC_approval), the land conversion of VA is rejected; otherwise, it is supported. This decision model is expressed as follows:

[0183]

[0184] Wherein, Econversion? represents the result of the EPC agent agreeing to or rejecting the land conversion, and EPC... approval For the approval rate of EPC agents;

[0185] Step 10.3: If EPC rejects VA's land conversion, VA repeats steps 8, 9 and 10 until the number of undeveloped pixels meets the target construction land size allocated by UDC to VA.

[0186] Step 11—Farmland Reclamation: The Farmland Conservation (FPC) agent selects a certain amount of land to be reclaimed as farmland based on its suitability for development. Farmland suitability is assessed using the same method as in Step 8.1. The amount of farmland reclaimed is determined by the total number of farmland pixels lost at each time step and the reclamation rate (FPC-reclaim, a preset value between 0 and 1). Specifically, it is represented as follows:

[0187]

[0188] In the formula, FPC_reclaim represents the amount of arable land reclaimed in the specified year t; FPC_reclaim is the arable land reclamation rate, with a value between 0 and 1. 0 indicates no arable land reclamation activity, representing the complete failure of the arable land occupation-compensation balance policy, while 1 indicates 100% implementation of the "occupy as much as you reclaim" arable land compensation policy. This represents the amount of arable land occupied by construction in the study area for the corresponding year.

[0189] Step 12: The Urban Construction Agent (UDC) calculates the total cumulative amount of newly added construction land in the city and makes a decision based on the CG decision model: If the total cumulative amount of newly added construction land exceeds the planning target set by the Central Government Agent (CG) for the planning period, the CG will adjust the approval rate for next year (a preset value of 0 to 1). If the CG approval rate is 0, the model stops. Specifically, it is expressed as follows:

[0190]

[0191] CG approval Represents the CG approval rate, count construction It is the cumulative amount of land converted into construction land starting from the initial year, quota construction This represents the planning indicators set by CG in the initial planning stage (Step 2). The model stops if the CG approval rate is 0.

[0192] The simulation period is extended to a planning period T (e.g., 15 years):

[0193] Step 13: The model stops, and the land use status is updated.

[0194] The effects of implementing the multi-level intelligent agent land change simulation method considering multiple policy interactions disclosed in this invention:

[0195] 1. Compared with existing research, this invention expands the study of urban land expansion mechanisms under multiple policy interactions: Based on the hierarchy of my country's planning system and the hierarchical management model of government land use, this invention provides a land change simulation method that couples multiple land policies and multi-level intelligent agent system interactions. This method helps to quantitatively and visually express the interaction between multi-level governments and policies behind urban land change (urban expansion), which is beneficial to subsequent subject modeling research based on policy interactions and mechanism research on human-land system coupling.

[0196] 2. Probability Assessment of Urban Construction Land Development: Natural factors included elevation, slope, vegetation cover index, and physical distance factors (distance to rivers and lakes, distance to tourist attractions, and distance to administrative centers—city centers, district centers, and township centers); socio-economic driving factors included GDP, population density, POI kernel density, distance to educational resources, distance to hospitals, and distance to transportation facilities; transportation factors included distance to airports, bus stations, train stations, and major roads. A random forest model was used to assess the probability of urban construction land development, with a model training accuracy of 0.91.

[0197] 3. The accuracy of the model proposed in this invention:

[0198] The proposed model was compared with a reference model (a cellular automaton model without coupled multiple policies and multi-level intelligent agent systems) using a unit-level consistency quality factor (FOM) metric to evaluate model accuracy.

[0199] (1) Evaluation of the overall accuracy of the model over the years during the study period

[0200] The FOM index value of the model proposed in this invention is greater than that of the reference model, which indicates that the proposed model has produced relatively high simulation results in terms of unit-level consistency, and that the model proposed in this invention can better replicate and "reproduce" the spatial pattern of urban land use change.

[0201] (2) Accuracy evaluation in each administrative region during the study period

[0202] The FOM index values ​​for simulated urban land use in each administrative region over the years were calculated using this model. The results show that the model proposed in this invention provides higher unit-level consistency in the rapidly growing central urban areas and urban hinterlands than in suburbs, and has better simulation accuracy in the central urban areas and urban hinterlands. This indicates that the model proposed in this invention has significant advantages in simulating urban land expansion, especially in simulating land use change hotspots driven by multiple policies.

[0203] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A land change simulation method considering multiple policy interactions and multi-level intelligent agents, characterized in that, Includes the following steps: S1: Construct a central government intelligent agent. Based on the planning data, use the central government intelligent agent to obtain planning indicators and central government approval rate parameters. The planning indicators include construction land planning indicators, cultivated land protection scale indicators, and ecological red line protection scale indicators. S2: Construct an intelligent agent for urban construction, and use the intelligent agent for urban construction to obtain the expected urban development needs; S3: Construct a village collective intelligent agent, and use the village collective intelligent agent to obtain the expected village development needs; S4: Based on the central government's approval rate, expected urban development needs, and expected village development needs, the urban construction agent allocates the annual construction land development scale to each village. S5: Based on the development scale of the construction land, land acquisition activities are carried out using the urban construction intelligence agent and the village collective intelligence agent to obtain a new development scale of village construction land; S6: Utilize the geospatial dataset to obtain the probability of construction land development and the initial land use status. Based on the scale of construction land development in the village, use the village collective agent to propose land conversion requests based on the probability of construction land development, generating the number of land conversions, including: Step 6.1—Initial Development Probability Assessment: Based on natural geographical, socio-economic, and transportation location factors, comprehensively assess the initial development probability of urban construction for each raster cell; Step 6.2—Local Development Probability Assessment, i.e. Neighborhood Effect Assessment: Considering the neighborhood land use effect, it is regarded as one of the important factors affecting the change of pixel status. The more neighborhood pixels around a certain pixel, the greater the probability that it will be converted into construction land. Step 6.3—Evaluation of the final development probability of construction land: The development probability of construction land is jointly determined by natural conditions, socio-economic factors and transportation location conditions, the neighborhood effect of construction land, and random factors, as expressed by the following formula: In the formula, For pixels ij The probability of the final development of the construction land; For pixels ij The initial probability of land development for construction is determined by factors including natural conditions, socio-economic factors, and transportation location conditions. For pixels ij The probability of local development, i.e., the neighborhood effect; It is a random number between 0 and 1, used to represent the uncertainty of land development for construction. , , It is the weight of the initial development probability, the local development probability, and the random probability, and ; S7: Construct a farmland protection intelligent agent and generate farmland conversion approval rate parameters; based on the land conversion request, use the farmland protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent. If the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to step S6. S8: Construct an ecological protection intelligent agent and generate ecological land conversion approval rate parameters; based on the land conversion request, use the ecological protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent. If the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to step S6. S9: Use the farmland protection intelligent agent to generate farmland reclamation rate parameters, obtain the amount of farmland converted and the amount of farmland reclaimed, and update the land use status; S10: Based on the land conversion quantity and land use status, the cumulative total amount of newly added construction land is obtained using the central government intelligent agent; based on the cumulative total amount of newly added construction land and planning indicators, a new central government approval rate is obtained; S11: Repeat steps S4-S10 to perform the simulation. When the number of land conversions is not less than the scale of construction land development, or the developable land of the village collective is less than the preset value, or the simulation year exceeds the last year of the preset simulation, the model stops and a new land use status is obtained.

2. The land change simulation method considering multiple policy interactions and multi-level intelligent agents as described in claim 1, characterized in that, Step S2 specifically includes: constructing an intelligent agent for urban development, and using the intelligent agent to obtain the expected urban development needs, as shown in the formula: in, This is the projected amount of municipal-level construction land. It is a predicted amount of land for construction based on trend extrapolation. It is based on the amount of construction land predicted by Markov, and y(t) is a curve function with respect to time t. and The coefficient of the variable. It is a constant term. and Represents time and Land use status, This represents the transition probability between land use types i and j.

3. The land change simulation method considering multiple policy interactions and multi-level intelligent agents as described in claim 1, characterized in that, Step S4 specifically includes: based on the central government approval rate, expected urban development needs, and expected village development needs, using the urban construction agent to allocate the annual construction land development scale for each village, as shown in the formula: In the formula, For each village collective intelligence entity, allocate construction land to the city's construction intelligence entity in a specified year t. For the approval rate, For a given year t, the predicted urban development demand is... For a given year t, the predicted development needs of the i-th village collective intelligence agent are... The sum of the expected needs of all villages. The approval rate represents the scale of land development allocated to each village collective intelligence by the city construction intelligence agent; the value of the approval rate is not less than 0 and not greater than 1.

4. The land change simulation method considering multiple policy interactions and multi-level intelligent agents as described in claim 1, characterized in that, Step S5 specifically includes: based on the development scale of the construction land, using the urban construction agent and the village collective agent to conduct land acquisition activities, obtaining the result of the village collective agreeing or refusing to sell the land, and obtaining a new village construction land development scale based on the development scale of the construction land and the result of the village collective agreeing or refusing to sell the land, as shown in the formula: in, The outcome of the village collective's decision to agree to or refuse to sell the land. Expressing approval, To indicate refusal, The probability that a village collective will refuse land acquisition by an urban construction entity. r The generated random number is a random number between [0 and 1]. r Greater than If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

5. A land change simulation system considering multiple policy interactions and a multi-level intelligent agent, characterized in that, Includes the following modules: The planning indicators module is configured as follows: Construct a central government intelligent agent, and based on the planning data, use the central government intelligent agent to obtain planning indicators and central government approval rate parameters. The planning indicators include construction land planning indicators, cultivated land protection scale indicators and ecological red line protection scale indicators. The module for generating anticipated urban development needs is configured to: construct an intelligent agent for urban construction, and use the intelligent agent for urban construction to obtain anticipated urban development needs; The module for generating expected village development needs is configured to: construct a village collective intelligent agent and use the village collective intelligent agent to obtain expected village development needs; The construction land development scale generation module is configured to: allocate the annual construction land development scale to each village using the urban construction agent based on the central government approval rate, expected urban development needs, and expected village development needs; The construction land development scale update module is configured to: based on the construction land development scale, use the urban construction agent and the village collective agent to carry out land acquisition activities to obtain a new village construction land development scale; The land conversion request module is configured to: obtain the development probability of construction land and the initial land use status using a geospatial dataset; based on the development scale of construction land in the village, use the village collective intelligent agent to propose land conversion requests based on the development probability of construction land, and generate the number of land conversions. The specific configuration is as follows: Initial development probability assessment: The initial development probability of urban construction for each raster cell is comprehensively assessed based on natural geographical, socio-economic, and transportation location factors. Local development probability assessment, also known as neighborhood effect assessment: Considering the neighborhood land use effect, it is regarded as one of the important factors affecting the change of pixel status. The more neighborhood pixels around a certain pixel, the greater the probability that it will be converted into construction land. Assessment of the final development probability of construction land: The development probability of construction land is jointly determined by natural conditions, socio-economic factors and transportation location conditions, the neighborhood effect of construction land and random factors, as expressed by the following formula: In the formula, For pixels ij The probability of the final development of the construction land; For pixels ij The initial probability of land development for construction is determined by factors including natural conditions, socio-economic factors, and transportation location conditions. For pixels ij The probability of local development, i.e., the neighborhood effect; It is a random number between 0 and 1, used to represent the uncertainty of land development for construction. , , It is the weight of the initial development probability, the local development probability, and the random probability, and ; The farmland protection decision-making module is configured to: construct a farmland protection intelligent agent and generate farmland conversion approval rate parameters; based on the land conversion request, use the farmland protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent; if the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to the land conversion request module. The ecological protection decision-making module is configured to: construct an ecological protection intelligent agent and generate ecological land conversion approval rate parameters; based on the land conversion request, use the ecological protection intelligent agent to make a behavioral decision on the land conversion request of the village collective intelligent agent; if the land conversion request is approved, update the land use status; if the land conversion request is rejected, return to the land conversion request module. The cultivated land reclamation quantity module is configured to: generate cultivated land reclamation rate parameters using the cultivated land protection intelligent agent, obtain the converted cultivated land quantity and cultivated land reclamation quantity, and update the land use status; The central government approval rate update module is configured to: obtain the cumulative total amount of newly added construction land based on the land conversion quantity and land use status using the central government intelligent agent; Based on the cumulative total amount of newly added construction land and planning indicators, a new approval rate from the central government is obtained; The iterative module is configured to repeat the simulation process of the construction land development scale generation module, construction land development scale update module, land conversion request module, farmland protection decision module, ecological protection decision module, farmland reclamation quantity module, and central government approval rate update module. When the land conversion quantity is not less than the construction land development scale, or the developable land of the village collective intelligence is less than the preset value, or the simulated year exceeds the last year of the preset simulation, the model stops and a new land use status is obtained.

6. The system according to claim 5, characterized in that, The specific configuration of the expected urban development demand generation module is as follows: Construct an intelligent agent for urban construction; utilize this intelligent agent to obtain the expected urban development demands, as shown in the formula: in, This is the projected amount of municipal-level construction land. It is a predicted amount of land for construction based on trend extrapolation. It is based on the amount of construction land predicted by Markov, and y(t) is a curve function with respect to time t. and The coefficient of the variable. It is a constant term. and Represents time and Land use status, This represents the transition probability between land use types i and j.

7. The system according to claim 5, characterized in that, The specific configuration of the construction land development scale generation module is as follows: based on the central government approval rate, expected urban development needs, and expected village development needs, the urban construction agent allocates the annual construction land development scale to each village, as shown in the formula: In the formula, For each village collective intelligence entity, allocate construction land to the city's construction intelligence entity in a specified year t. For the approval rate, For a given year t, the predicted urban development demand is... For a given year t, the predicted development needs of the i-th village collective intelligence agent are... The sum of the expected needs of all villages. The approval rate represents the scale of land development allocated to each village collective intelligence by the city construction intelligence agent; the value of the approval rate is not less than 0 and not greater than 1.

8. The system according to claim 5, characterized in that, The specific configuration of the construction land development scale update module is as follows: Based on the construction land development scale, land acquisition activities are carried out using the urban construction agent and the village collective agent to obtain the result of the village collective agreeing or refusing to sell the land. Based on the construction land development scale and the result of the village collective agreeing or refusing to sell the land, a new village construction land development scale is obtained, as shown in the formula: in, The outcome of the village collective's decision to agree to or refuse to sell the land. Expressing approval, To indicate refusal, The probability that a village collective will refuse land acquisition by an urban construction entity. r The generated random number is a random number between [0 and 1]. r Greater than If the village collective intelligence agrees to the urban construction intelligence's land acquisition, then the village collective intelligence will agree to the urban construction intelligence's land acquisition; otherwise, the village collective intelligence will refuse the urban construction intelligence's land acquisition, and the village collective intelligence's construction land development scale for this year will be set to 0.

Citation Information

Patent Citations

  • Multi-agent urban expansion simulation method based on land acquisition process competition result

    CN104318319A

  • Town settlement space development digital simulation method and system

    CN112651141A