Adaptive planning partitioning method based on improved random seed and simulated annealing

Abstract

The present invention discloses an adaptive planning partitioning method based on improved random seed and simulated annealing, comprising the following steps: initial data processing; based on density analysis, selecting a high-dominance map as a candidate seed point; determining the appropriate partition number interval of the study area; looping through each partition number ZoneNum, performing simulated annealing, setting the initial temperature, iteratively updating and updating to find the optimal solution, and obtaining the evaluation function value under different partition numbers; after traversing all partition numbers, the change of the evaluation function value under different partition numbers is plotted into a line graph, a curve of the effect evaluation index changing with the number of partitions is constructed, and the partition scheme at the inflection point of the curve is found to be the optimal planning partition scheme. The method uses an artificial intelligence algorithm to combine the KD-tree clustering method with simulated annealing to achieve adaptive partitioning from the land use results at the plot level.

Description

Technical Field

[0001] The present invention belongs to the field of national land space planning zoning, and specifically relates to an adaptive planning zoning method based on improved random seeds and simulated annealing. Background Art

[0002] At present, the zoning of national land space planning is a necessary means to optimize the functional layout structure and national land space pattern. The purpose of determining the planning zoning is to undertake the main functional zoning upwards, and at the same time radiate downwards to guide the specific management and control of land use, forming a structural and functional regional division system to supplement and improve the existing spatial zoning system, and further clarify the functional positioning and land use control principles of each region through scientific and reasonable division of regional space, so as to achieve the optimal allocation of resources and maximize the efficiency of use.

[0003] In order to ensure the scientificity and effectiveness of planning zoning, basic planning zoning principles must be followed. First, the integrity of administrative boundaries must be guaranteed, that is, each administrative division must clearly define its jurisdiction, and no intersection or overlap is allowed; second, the unity of functions, the functional layout in different regions must be coordinated with each other to avoid functional conflicts and duplication of construction; finally, the connection and merger of space, that is, close links must be formed between regions, and spatial integration and integrated construction within large regions must be strengthened. In actual operation, the division of national land space planning zoning needs to comprehensively consider geographical, social, economic and other factors, effectively guide and regulate land use behavior, prevent disorderly expansion and waste of resources, and create a more livable living environment for urban and rural residents, and promote my country to achieve high-quality regional coordinated development goals.

[0004] The zoning of national land space planning is a complex and meticulous task, which involves comprehensive consideration of many factors such as land resources, ecological environment, and economic development.

[0005] At present, based on existing research and practical experience, the partitioning methods can be roughly divided into three types:

[0006] First, according to the zoning objectives, an indicator system of resource and environmental carrying capacity evaluation, ecosystem services, etc. is constructed, and the types of preset units are identified by using methods such as principal component analysis and expert decision-making. This method is mostly used at the macro scale, ignoring the differences in zoning types at different scales, resulting in the inability to implement planning zoning boundaries at the plot level, and the lack of consistency and comparability of zoning results at different levels;

[0007] Second, based on comparable units such as grids, land use patterns, and parcels, conduct a comprehensive analysis and evaluation of the single and mixed functions of zoning units on a spatial basis, continuously combine and determine zoning results, and compare zoning results between comparable units to improve the accuracy of zoning delineation;

[0008] 3. Based on the spatial association theory, the partition units are divided according to the strength of the spatial relationship. Among them, spectral clustering, K-means clustering and Q clustering based on graph theory are often used to reveal the structural relationship between spatial units and achieve efficient partitioning.

[0009] However, these methods are relatively slow when processing large data sets or high-dimensional data due to their high computational complexity, and their computational efficiency needs to be weighed in practical applications.

[0010] Therefore, there is an urgent need for a method that can combine artificial intelligence algorithms to achieve adaptive zoning from land use results at the plot level. Summary of the invention

[0011] In response to the above technical problems, the present invention proposes a zoning delineation method that uses an artificial intelligence algorithm to combine the KD-tree (tree data structure) clustering method with simulated annealing, to achieve adaptive zoning delineation from land use results at the plot level, and then to achieve fine and reasonable zoning of national land space planning at the plot scale.

[0012] To achieve the above purpose, the technical solution provided by the present invention is as follows:

[0013] In a first aspect, an adaptive planning partitioning method based on improved random seed and simulated annealing is provided, comprising the following steps:

[0014] Obtain the current land use status map, couple the dual evaluation and planning constraints, and use the vector CA simulation method to generate the future planned land use layout as the starting data for planning zoning;

[0015] The dominance of the regional dominant function is evaluated by density analysis of the initial data of the planning zone, and the patches with high dominance are selected as candidate seed points;

[0016] Determine the appropriate zone number interval [MinZoneNum, MaxZoneNum] for the planned zone, where the minimum zone number MinZoneNum is the number of dominant use types; the maximum zone number MaxZoneNum is the number of all candidate seed points;

[0017] Loop through the appropriate partition number interval, perform simulated annealing for each partition number ZoneNum, set the initial temperature, iterate and update until the maximum number of times, and obtain the evaluation function value under different partition numbers;

[0018] After traversing all the numbers of partitions, the changes in the evaluation function values ​​under different numbers of partitions are plotted as a line graph, and a function curve of the effect evaluation index changing with the number of partitions is constructed. The partition scheme at the inflection point of the curve is found to be the optimal planning partition scheme.

[0019] In a possible implementation, the specific formula for coupling the dual evaluation and planning constraints and using the vector CA simulation method to predict the future planned land use layout is as follows:

[0020]

[0021] in Is the plot at time Change to other land use types The possibility of is a random disturbance term, It is a plot of land accessibility, It is a plot of land Land use type The suitability of It's time Land Use Plots With distance The domain effect between adjacent parcels within the buffer zone, For plots The double evaluation result parameters, For plots planning constraints.

[0022] In a possible implementation manner, the starting data of the planning zone is evaluated for the dominance of the regional dominant function by means of density analysis, wherein the dominance calculation method is shown in the following formula:

[0023]

[0024] in, express Plots Dominance of land use types; express The land use types in all rasters contained in the plot The ratio of the number of grids; a It represents the weight, which varies according to the land use type and is determined by the Delphi method.

[0025] In a possible implementation manner, the simulated annealing method is as follows:

[0026] The seed points are selected randomly under conditions, a multidimensional tree data structure (KD-tree) is constructed to generate the initial planning partition scheme A0, the objective function value F(A0) corresponding to the solution A0 is calculated, and the initial temperature T is set;

[0027] The perturbation randomly generates a new planning partition scheme A1, calculates the objective function value F(A1) corresponding to the solution A1, determines whether to accept the new scheme based on the Metropolis criterion, and updates the current temperature until the maximum number of iterations is reached;

[0028] In an improved implementation, the conditionally randomly selects seed points, constructs a multidimensional KD-tree to generate an initial planning partitioning scheme A0, calculates the objective function value F(A0) corresponding to the solution A0, and sets the initial temperature T, including:

[0029] ZoneNum seed points are randomly selected from the candidate seed points under certain conditions. The conditions include: (1) the random seeds are from the candidate seeds with dominance constraints; (2) at least one seed is guaranteed for each partition type, and each seed has at least three attributes: geographic coordinates, area, and planned land use type;

[0030] Using geographic coordinates, area, and planned land use type as key attributes, a multidimensional KD-tree is constructed for clustering. Using the KD-Tree query function, all plots are traversed and classified into the partitions represented by each seed point to form a planned partition scheme, i.e., a new solution.

[0031] The entropy value within each partition unit and the inter-group entropy value between different partition units are calculated to construct an evaluation function integrating the land use entropy value, which is the objective function value F(A0). The formula is as follows:

[0032]

[0033]

[0034]

[0035]

[0036] in, is the evaluation function, , Refers to the number of subdivisions and the total number of plots, respectively. E Refers to information entropy, Refers to the sum of information entropy of each partition, It refers to the ratio of the total area of ​​each land use type in the zone to the total area of ​​the zone. Refers to the total information entropy based on the area of ​​each partition and the dominant land type within the partition unit. It refers to the ratio of the sum of the areas of the dominant land types in a subdivision unit to the total area of ​​the subdivision unit. is the proportionality factor, which is used to eliminate the comparability differences caused by the different number of partitions. lg () is the logarithm with base 10.

[0037] An improved implementation method is as follows: randomly generate a new planned partition scheme A1 by perturbation, calculate the objective function value F(A1) corresponding to the solution A1, judge whether to accept the new scheme based on the Metropolis criterion, and update the current temperature until the maximum number of iterations is reached:

[0038] (1) Randomly generate a new planned partition scheme A1 by perturbation; among them, the method for selecting seeds is: based on the current optimal partition scheme A0, traverse each planned partition, and randomly select 1 homogeneous patch as a new seed from the candidate seeds within the planned partition, ensuring that the result of the original A0 is retained to the greatest extent, while having random perturbation. After the traversal is completed, ZoneNum seeds can be formed;

[0039] (2) Calculate the objective function value F(A1) corresponding to the solution A1 based on the calculation formula of the evaluation function:

[0040] Judge whether to accept the new scheme according to the Metropolis criterion: if F(A1)>F(A0), then assign the solution A1 to the solution A0 and calculate the objective function value; if F(A1)≤F(A0), calculate the acceptance probability p of A1, and then generate a random number r between [0,1]. If r<p, assign the solution A1 to the solution A0 and calculate the objective function value; otherwise, directly calculate the objective function value; during the process, the temperature decreases as the number of iterations t increases, so as to reduce the probability of accepting inferior solutions and tend to the global optimal solution. The specific formula of the acceptance probability function is as follows.

[0041]

[0042] Among them, is the acceptance function, which refers to the probability of replacing the existing optimal solution with the new optimal solution after the result of a new iteration is generated; is the interpolation between the evaluation value of the new iteration result and the existing optimal solution. A negative value represents that the effect of the new iteration result is better than the existing optimal solution, and vice versa; is the number of iterations.

[0043] Judge whether to end the iteration: whether the number of iterations reaches the maximum value. If not, continue the iteration; otherwise, obtain the optimal planned partition corresponding to the current partition number ZoneNum and the corresponding optimal evaluation function value.

[0044] In the second aspect, an adaptive planning partition device based on improved random seeds and simulated annealing is provided, including:

[0045] An acquisition module, configured to acquire the patches of the existing land use status, couple the dual evaluation and planning constraint conditions, and use the simulation method of vector CA to generate the future planned land use layout as the starting data of the planning partition;

[0046] The selection module is used to evaluate the dominance of the regional dominant function by density analysis of the starting number of the planning partitions, and select the patches with high dominance as candidate seed points;

[0047] The module for determining the number of zones is used to determine the appropriate zone number interval [MinZoneNum, MaxZoneNum] for the planned zone, where the minimum zone number MinZoneNum is the number of dominant use types; the maximum zone number MaxZoneNum is the number of all candidate seed points;

[0048] The simulated annealing module is used to loop through the interval of suitable partition numbers, perform simulated annealing for each partition number ZoneNum, set the initial temperature, iterate and update until the maximum number of times, and obtain the evaluation function value under different partition numbers;

[0049] The scheme acquisition module is used to traverse all the numbers of partitions, draw the changes in the evaluation function values ​​under different numbers of partitions into a line graph, construct a function curve of the effect evaluation index changing with the number of partitions, and find the partition scheme at the inflection point of the curve as the optimal planning partition scheme.

[0050] In a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the adaptive planning partitioning method based on improved random seeds and simulated annealing as described above is implemented.

[0051] In a fourth aspect, a non-transitory computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the adaptive planning partitioning method based on improved random seeds and simulated annealing as described above.

[0052] In a fifth aspect, a computer program product comprises a computer program, which, when executed by a processor, implements the adaptive planning partitioning method based on improved random seeds and simulated annealing as described above.

[0053] The planning zoning method proposed in the present invention has high application value. Compared with the existing planning zoning method, the advantages and beneficial effects of the present invention are embodied in the following aspects:

[0054] First, the present invention proposes a method for evaluating the dominance of dominant functions. Based on the existing land use status map, the vector CA simulation method is used, dual evaluation and other planning constraints are coupled, and the dominance of regional dominant functions is evaluated by density analysis. Finally, the main functional types of zoning units are accurately identified by combining the existing permanent basic farmland and ecological protection red line data.

[0055] Second, the present invention innovatively improves the random seed algorithm. On the basis of determining the appropriate partition number interval of the study area, seed points are conditionally randomly selected from the candidate seeds to ensure that at least one seed point is selected for each partition type.

[0056] Third, the present invention creatively constructs a Calinski-Harabasz index that integrates the improved land use entropy value as an optimization utility function, which effectively evaluates the partitioning results in the iterative process and improves the scientificity and practicality of the partitioning results.

[0057] Fourthly, the planning partition optimization process of the present invention adopts an artificial intelligence algorithm to avoid local optimal solutions and find the global optimal partitioning scheme through multiple iterative optimizations.

[0058] Fifth, the present invention applies the KD-Tree method to perform cluster analysis in planning partitions, which is suitable for processing large-scale data and improves the calculation efficiency of planning partitions. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] FIG1 is a schematic diagram of a flow chart of an adaptive planning partitioning method based on random seeds and simulated annealing according to the present invention;

[0060] FIG2 is a curve relationship diagram of the number of partitions and the partition evaluation function value in the present invention;

[0061] Figure 3 A schematic diagram of the structure of the adaptive planning partitioning device based on improved random seeds and simulated annealing provided by the present invention;

[0062] FIG. 4 is a schematic diagram of the structure of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0063] The present invention is further elaborated in the following with reference to a specific example of a city's land space planning zoning, but the content of the present invention is not limited thereto.

[0064] like Figure 1 As shown, an adaptive planning partitioning method based on random seeds and simulated annealing includes the following steps:

[0065] S100 obtains the current land use status map, couples the dual evaluation and planning constraints, and uses the vector CA (cellular automata) simulation method to generate the future planned land use layout as the starting data for planning zoning.

[0066] In a specific embodiment, the specific formula for coupling the dual evaluation and planning constraints and using the vector CA simulation method to predict the future planned land use layout is as follows:

[0067]

[0068] in Is the plot at time Change to other land use types The possibility of is a random disturbance term, It is a plot of land accessibility, It is a plot of land Land use type The suitability of It's time Land use type Plots With distance The domain effect between adjacent parcels within the buffer zone, For plots The double evaluation result parameters, For plots According to the planning constraints, corresponding values ​​are taken for the forest land, grassland and water area within the ecological protection red line, the cultivated land within the permanent basic farmland, the national ecological protection area, the coastal mangrove area, and some slender features such as roads and rivers that are easily ignored in the raster model. The remaining plots and types are assigned a value of 0 to ensure that the land use type in important areas does not change.

[0069] For example, based on the land use changes in a certain city in 2010 and 2020, the vector CA model is used to predict the land use in 2030, and the land use raster simulation results of the city in 2030 are obtained as the starting data for planning zoning.

[0070] S200, evaluating the dominance of the regional dominant function by density analysis of the initial data of the planning partition, and selecting the patches with high dominance as candidate seed points.

[0071] In a specific embodiment, the predicted simulated future land use layout is evaluated by density analysis to determine the dominance of the regional dominant function, wherein the dominance calculation method is as shown in the following formula:

[0072]

[0073] in, express Plots Dominance of land use types; express The land use types in all rasters contained in the plot The ratio of the number of grids; aIt represents weight, and its value varies according to the land use type. It is determined by the Delphi method. In this example, the value of cultivated land is 0.4, the value of garden land is 0.1, the value of forest land is 0.2, and the value of construction land is 0.3.

[0074] S300, determining the appropriate zone number interval [MinZoneNum, MaxZoneNum] for the study area, where the minimum zone number MinZoneNum is the number of dominant use types; the maximum zone number MaxZoneNum is the number of all candidate seed points.

[0075] For example, in this embodiment, the interval of the determined number of zones is [4,20]. That is, MinZoneNum is 4, indicating that the number of dominant use types is 4, namely, cultivated land, forest land, garden land, and construction land; and MaxZoneNum is 20.

[0076] S400, looping through the appropriate partition number interval, performing simulated annealing for each partition number ZoneNum, setting the initial temperature, iterating and updating until the maximum number of times, and obtaining the evaluation function values ​​under different partition numbers.

[0077] In a specific embodiment, S400 includes the following sub-steps:

[0078] S410, conditionally randomly select seed points, construct a multi-dimensional KD-tree (tree data structure) to generate an initial planning partition scheme A0, calculate the objective function value F(A0) corresponding to the solution A0, and set the initial temperature T;

[0079] S420, randomly generate a new planning partition scheme A1 by disturbance, calculate the objective function value F(A1) corresponding to the solution A1, determine whether to accept the new scheme based on the Metropolis criterion, and update the current temperature until the maximum number of iterations is reached;

[0080] S430, obtaining evaluation function values ​​under different numbers of partitions.

[0081] Specifically, the method of S410 includes:

[0082] S411, randomly select ZoneNum seed points from the candidate seed points conditionally; the conditions include: (1) the random seeds are from the candidate seeds with dominance constraints; (2) at least one seed is guaranteed for each partition type, and each seed has at least three attributes, namely, geographic coordinates (latitude and longitude), area, and planned land use type;

[0083] S412, using geographic coordinates, area, and planned land use type as key attributes, construct a multidimensional KD-tree for clustering; using the KD-Tree query function, traverse all plots and classify them into the partitions represented by each seed point to form a planned partition scheme, that is, a new solution;

[0084] S413, calculate the entropy value within each partition unit and the inter-group entropy value between different partition units, and construct the Calinski-Harabasz index (evaluation function) improved by integrating the land use entropy value, which is the objective function value F(A0), and the formula is as follows:

[0085]

[0086]

[0087]

[0088]

[0089] in, is the evaluation function, , Refers to the number of subdivisions and the total number of plots, respectively. E Refers to information entropy, Refers to the sum of information entropy of each partition, It refers to the ratio of the total area of ​​each land use type in the zone to the total area of ​​the zone. Refers to the total information entropy based on the area of ​​each partition and the dominant land type within the partition unit. It refers to the ratio of the sum of the areas of the dominant land types in a subdivision unit to the total area of ​​the subdivision unit. is the proportionality factor, which is used to eliminate the comparability differences caused by the different number of partitions. lg () is the logarithm with base 10.

[0090] Specifically, the method of S420 includes:

[0091] S421, randomly generate a new planning partition scheme A1 by disturbance; wherein, the method of selecting seeds is: based on the current optimal partition scheme A0, traverse each planning partition, randomly select a similar map spot from the candidate seeds in the planning partition as a new seed, ensure that the original A0 result is retained to the maximum extent, and at the same time, there is random disturbance, and after the traversal is completed, ZoneNum seeds can be formed;

[0092] S421, based on the calculation formula of the evaluation function, calculate the objective function value F(A1) corresponding to the solution A1:

[0093] Judge whether to accept the new solution according to the Metropolis criterion: If F(A1) > F(A0), then assign the solution A1 to the solution A0 and calculate the objective function value; if F(A1) ≤ F(A0), calculate the probability p of accepting A1, then generate a random number r between [0, 1]. If r < p, assign the solution A1 to the solution A0 and calculate the objective function value; otherwise, directly calculate the objective function value. During the process, the temperature decreases as the number of iterations t increases, so as to reduce the probability of accepting inferior solutions and tend to the global optimal solution. The specific formula of the acceptance probability function is as follows.

[0094]

[0095] Among them, is the acceptance function, which refers to the probability of replacing the existing optimal solution with the new optimal solution after the result of a new iteration is generated; is the interpolation between the evaluation value of the new iteration result and the existing optimal solution. A negative value represents that the result of the new iteration is better than the existing optimal solution, and vice versa; is the number of iterations.

[0096] S422, judge whether to end the iteration: Whether the number of iterations reaches the maximum value (2000 times). If not, continue the iteration. Otherwise, obtain the optimal planning partition corresponding to the current number of partitions ZoneNum and the corresponding optimal evaluation function value, that is, the optimal objective function value.

[0097] S500, after traversing all the numbers of partitions, plot the change of the evaluation function values under different numbers of partitions as a line chart, construct a function curve of the effect evaluation index changing with the number of partitions, and find the partition scheme at the inflection point of the curve as the best planning partition scheme. As Figure 2 shown, the evaluation function value reaches the highest when the number of partitions in this study area is 6 at the inflection point of the curve.

[0098] Based on the partition results, the partition units smaller than 1 km 2 are aggregated into the surrounding other partition patches according to the largest area, and combined with the administrative region boundary, permanent basic farmland, and ecological protection red line data. The partitions with the planned dominant functions of cultivated land and garden land and containing contiguous permanent basic farmland are set as the agricultural core areas, and those without contiguous permanent basic farmland are set as rural development areas; for those with the dominant land use functions of forest land and grassland, those containing contiguous ecological protection red lines are set as ecological core areas, and vice versa as ecological control areas; those with the dominant function of construction land are set as urban development zones; the coastal areas are set as marine functional zones. The area of each partition is shown in Table 1.

[0099] Table 1

[0100]

[0101] The following is a description of the adaptive planning partitioning device based on improved random seeds and simulated annealing provided by the present invention. The adaptive planning partitioning device based on improved random seeds and simulated annealing described below and the adaptive planning partitioning method based on improved random seeds and simulated annealing described above can be referred to each other.

[0102] Figure 3 is a schematic diagram of the structure of an adaptive planning partitioning device based on improved random seeds and simulated annealing provided by an embodiment of the present invention, such as Figure 3 As shown, it includes: an acquisition module 31, a selection module 32, a partition number determination module 33, a simulated annealing module 34 and a solution acquisition module 35, wherein:

[0103] The acquisition module 31 is used to obtain the current land use status map, couple the dual evaluation and planning constraints, and use the vector CA simulation method to generate the future planned land use layout as the starting data for the planning zoning;

[0104] A selection module 32 is used to evaluate the dominance of the regional dominant function by density analysis of the initial data of the planning partition, and select the patches with high dominance as candidate seed points;

[0105] The partition number determination module 33 is used to determine the appropriate partition number interval [MinZoneNum, MaxZoneNum] of the planned partition, wherein the minimum partition number MinZoneNum is the number of dominant use types; the maximum partition number MaxZoneNum is the number of all candidate seed points;

[0106] The simulated annealing module 34 is used to loop through the interval of suitable partition numbers, perform simulated annealing for each partition number ZoneNum, set the initial temperature, iterate and update until the maximum number of times, and obtain the evaluation function value under different partition numbers;

[0107] The scheme acquisition module 35 is used to traverse all the numbers of partitions, plot the changes in the evaluation function values ​​under different numbers of partitions into a line graph, construct a function curve of the effect evaluation index changing with the number of partitions, and find the partition scheme at the inflection point of the curve as the optimal planning partition scheme.

[0108] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the adaptive planning partitioning method based on the improved random seed and simulated annealing.

[0109] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0110] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the adaptive planning partitioning method based on improved random seeds and simulated annealing provided by the above methods.

[0111] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the adaptive planning partitioning method based on improved random seeds and simulated annealing provided by the above-mentioned methods.

[0112] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0113] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive planning partitioning method based on improved random seeds and simulated annealing, characterized in that: The following steps are involved: Obtain the current land use status map, couple the dual evaluation and planning constraints, and use the vector CA simulation method to generate the future planned land use layout as the starting data for planning zoning; The dominance of the regional dominant function is evaluated by density analysis of the initial data of the planning zone, and the patches with high dominance are selected as candidate seed points; Determine the appropriate zone number interval [MinZoneNum, MaxZoneNum] for the planned zone, where the minimum zone number MinZoneNum is the number of dominant use types; The maximum number of zones MaxZoneNum is the number of all candidate seed points; Loop through the appropriate partition number interval, perform simulated annealing for each partition number ZoneNum, set the initial temperature, iterate and update until the maximum number of times, and obtain the evaluation function value under different partition numbers; The simulated annealing method is as follows: conditionally randomly select seed points, construct a multidimensional tree data structure to generate an initial planning partitioning scheme A0, calculate the objective function value F(A0) corresponding to the solution A0, and set the initial temperature T; randomly generate a new planning partitioning scheme A1 by perturbation, calculate the objective function value F(A1) corresponding to the solution A1, determine whether to accept the new scheme based on the Metropolis criterion, and update the current temperature until the maximum number of iterations is reached; In the simulated annealing method, a seed point is conditionally randomly selected, a multidimensional tree data structure is constructed to generate an initial planning partitioning scheme A0, an objective function value F(A0) corresponding to the solution A0 is calculated, and a method for setting an initial temperature T includes: ZoneNum seed points are randomly selected from the candidate seed points under certain conditions; the conditions include: (1) Random seeds come from candidate seeds with dominance constraints; (2) Each zoning type must have at least one seed, and each seed has at least three attributes: geographic coordinates, area, and planned land use type; Using geographic coordinates, area, and planned land use type as key attributes, a multidimensional tree data structure is constructed for clustering; using the tree data structure query function, all plots are traversed and classified into the partitions represented by each seed point to form a planned partition scheme, i.e., a new solution; Calculate the entropy value within each partition unit and the inter-group entropy value between different partition units, and construct the evaluation function number of the integrated land use entropy value, which is the objective function value F(A0), and the formula is as follows: in, is the evaluation function, , Refers to the number of subdivisions and the total number of plots, respectively. E Refers to information entropy, Refers to the sum of information entropy of each partition, It refers to the ratio of the total area of ​​each land use type in the zone to the total area of ​​the zone. Refers to the total information entropy based on the area of ​​each partition and the dominant land type within the partition unit. It refers to the ratio of the sum of the areas of the dominant land types in a subdivision unit to the total area of ​​the subdivision unit. is the proportionality factor, which is used to eliminate the comparability differences caused by the different number of partitions. lg () is the logarithm with base 10; After traversing all the numbers of partitions, the changes in the evaluation function values ​​under different numbers of partitions are plotted as a line graph, and a function curve of the effect evaluation index changing with the number of partitions is constructed. The partition scheme at the inflection point of the curve is found to be the optimal planning partition scheme.

2. The adaptive planning partitioning method based on improved random seed and simulated annealing according to claim 1, characterized in that: The specific formula for coupling the dual evaluation and planning constraints and using the vector CA simulation method to predict the future planning land use layout is as follows: in is the plot at time Change to other land use types The possibility of is a random disturbance term, It is a plot of land accessibility, It is a plot of land Land use type The suitability of It's time Land use type Plots With distance The neighborhood effect between adjacent parcels within the buffer zone of For plots The double evaluation result parameters, For plots planning constraints.

3. The adaptive planning partitioning method based on improved random seed and simulated annealing according to claim 1 is characterized in that: The initial data of the planning zoning is evaluated by density analysis to evaluate the dominance of the regional dominant function, wherein the dominance calculation method is shown in the following formula: in, express Plots Dominance of land use types; express The land use types in all rasters contained in the plot The ratio of the number of grids; a It represents the weight, which varies according to the land use type and is determined by the Delphi method.

4. The adaptive planning partitioning method based on improved random seed and simulated annealing according to claim 1 is characterized in that: The method of randomly generating a new planned partition scheme A1 by perturbation, calculating the objective function value F(A1) corresponding to the solution A1, judging whether to accept the new scheme based on the Metropolis criterion, and updating the current temperature until the maximum number of iterations is reached is as follows: (1) Perturbation randomly generates a new planned partition scheme A1; among them, the method of selecting seeds is as follows: based on the current optimal partition scheme A0, traverse each planned partition, and randomly select 1 homogeneous patch as a new seed from the candidate seeds within the planned partition, ensuring that the result of the original A0 is retained to the greatest extent, and at the same time performing random perturbation. After the traversal is completed, ZoneNum seeds can be formed; (2) Based on the calculation formula of the evaluation function, calculate the objective function value F(A1) corresponding to the solution A1: Judge whether to accept the new scheme according to the Metropolis criterion: if F(A1)>F(A0), then assign the solution A1 to the solution A0 and calculate the objective function value; if F(A1)≤F(A0), calculate the probability p of accepting A1, and then generate a random number r between [0,1]. If r<p, assign the solution A1 to the solution A0 and calculate the objective function value; otherwise, directly calculate the objective function value; during the process, the temperature decreases as the number of iterations t increases, so as to reduce the probability of accepting inferior solutions and tend to the global optimal solution. The specific formula of the acceptance probability function is as follows: in, is the acceptance function, which refers to the probability of replacing the existing optimal solution and becoming the new optimal solution after a new iteration result is generated; It is the interpolation between the evaluation value of the new iteration result and the existing optimal solution. If it is negative, it means that the new iteration result is better than the existing optimal solution, otherwise it means it is worse. is the number of iterations; Judge whether to end the iteration: whether the number of iterations reaches the maximum value. If not, continue the iteration; otherwise, obtain the optimal planned partition corresponding to the current number of partitions ZoneNum and the corresponding optimal evaluation function value.

5. An adaptive planning partitioning device based on improved random seed and simulated annealing, characterized in that: The method for implementing any one of claims 1-4 includes: An acquisition module, configured to acquire the patches of the existing land use status, couple the double evaluation and planning constraint conditions, and use the simulation method of vector CA to generate the future planned land use layout as the starting data of the planned partition; A selection module, configured to evaluate the dominant function superiority degree of the area by density analysis of the starting data of the planned partition, and select the patches with high superiority degree as candidate seed points; A partition number determination module, configured to determine the appropriate partition number interval [MinZoneNum, MaxZoneNum] of the planned partition, where the minimum partition number MinZoneNum is the number of dominant use types; the maximum partition number MaxZoneNum is the number of all candidate seed points; A simulated annealing module, configured to loop through the appropriate partition number interval, perform simulated annealing on each partition number ZoneNum, set the initial temperature, and iterate and update until the maximum number of times to obtain the evaluation function values under different partition numbers; A scheme acquisition module, configured to traverse all partition numbers, plot the change of the evaluation function values under different partition numbers as a line chart, construct a function curve of the effect evaluation index changing with the partition number, and find the partition scheme at the inflection point of the curve as the best planned partition scheme.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the adaptive planning partition method based on improved random seeds and simulated annealing as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the adaptive planning partitioning method based on improved random seeds and simulated annealing as claimed in any one of claims 1 to 4 is implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the adaptive planning partitioning method based on improved random seeds and simulated annealing as claimed in any one of claims 1 to 4 is implemented.

Images