Method for dividing renewable resource recycling area by identifying aggregation unit and barycenter area

Abstract

The present application relates to a kind of k piece division method of identifying aggregation unit and work focus area.The scheme adopts computer simulation division, first set the relevant data in the scope of project, and construct simulation matrix map by riding distance.In the running process, the originally K clustering division, based on the set sub-piece area unit coefficient ρ, original point position is scattered into K*ρ aggregation sub-unit and H discrete piece area unit;And based on the barycenter unit constantly absorbs, combine the non-barycenter unit of edge, until the piece area demand of integrated aggregation sub-unit is greater than or equal to the man efficiency of this piece area.Based on the piece area division result K+H finally distributed by such scheme, K is the number of aggregation sub-unit, H is the number of discrete piece area unit.In this way, the responsible piece area of collector can be reasonably divided, on the one hand, ensure that the workload of collector is basically consistent, ensure fairness;On the other hand, as far as possible, cover each community, ensure that the number of remaining discrete units is minimum.

Description

Technical Field

[0001] This invention relates to the field of project operation, and in particular to a method for dividing recycling zones for renewable resources that identifies clustered units and key areas. Background Technology

[0002] While the recycling and reuse of waste services is booming, a major challenge for environmental organizations and business operations management is how to balance the efficient use of collection personnel to improve the resident experience, ensure relative fairness in their income, and maintain a comfortable work environment. In reality, collection projects often cover vast areas with complex geographical environments, numerous locations with inconsistent daily efficiency, and require consideration of complex, multi-faceted objectives. Therefore, a digitalized area division method is urgently needed to rationally allocate work areas and assist managers in planning daily work area operations. Thus, researching work area allocation methods under complex objectives and scenarios is crucial for solving the challenges of managing collection personnel. Summary of the Invention

[0003] To address the aforementioned problems, the present invention aims to provide a method for dividing recycling areas into zones that identify clustered units and key regions. This method can reasonably divide the areas of responsibility for collectors, ensuring that the workload of collectors is basically consistent and thus ensuring fairness. On the other hand, it can cover all areas as much as possible to ensure that the number of remaining discrete units is minimized.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A method for identifying clustered units and key areas in the delineation of recycling zones for renewable resources, characterized by the following steps:

[0006] Step 1: Obtain the number of collectors K, the number of equipment points N, and the average daily delivery volume w of the i-th equipment. i Equipment location geographical location (longitude and latitude), geographical isolation threshold O, sub-area unit coefficient ρ, calculate per capita total area demand.

[0007] Step 2: Based on the geographical location (longitude and latitude) of the device points, obtain the distance d between each pair of points through the cycling distance API interface of the map software. ij The data is then processed and integrated into a cell distance matrix D to construct a simulated matrix map;

[0008] Step 3: The original points on the cell distance matrix map are broken down into K*ρ clustered sub-units and H discrete area units using the K-means clustering algorithm;

[0009] Step 4: Use the K-means clustering algorithm to select and divide the original points into K clustered sub-units, obtain the centroid of the K clustered sub-units, and substitute the centroid of the sub-units into the K*ρ clustered sub-units in Step 3. The clustered unit where the centroid point is located is the centroid unit.

[0010] Step 5: Using the centroid unit as the core, and following a rule from smallest to largest, match and combine other clustered units within the shortest distance O constraint to form regions. Continuously update the region with the smallest location until the demand of all regions is greater than or equal to W. avg +5000*coef [coef takes the value 0 or 1, and is defined as a parameter for whether to relax the constraints, 1 for relaxation and 0 for constraint], or when all non-centroidal units have been allocated or when they cannot be allocated due to exceeding the constraints, stop; resulting in K clustered sub-units and multiple discrete patch units;

[0011] Step 6: Calculate the discrete points globally and supplement the discrete patch units divided in Step 5 to obtain H discrete patch units;

[0012] Step 7: Repeat steps 3-6, and compare the result K+H of each area division, where K is the number of clustered sub-units and H is the number of discrete area units; take the result with the fewest remaining discrete units and the most balanced demand in the main area as the optimal area division result.

[0013] This invention employs the aforementioned technical solution, which relates to a method for dividing recyclable resource recovery zones by identifying clustered units and key areas. This method uses computer simulation for division, first setting the equipment points (recycling bin locations) within the project's scope, and the average daily delivery volume w at each equipment point. i And the number of collectors the project plans to deploy, to calculate the total demand per capita in the area.

[0014] Then, based on the GPS location of the device points in map software (such as Amap), the cycling distance between the device points is determined, thereby constructing a simulated matrix map; subsequent calculations and iterations are all performed on the simulated matrix map, thereby reducing the amount of computation.

[0015] During operation, considering that k-means has strong clustering identification capabilities but does not inherently consider demand balancing, a special design was implemented. The original K-clustering partitioning, based on a set sub-region unit coefficient ρ, broke down the original data points into K*ρ clustered sub-units and H discrete region units. Then, based on a set K-person partitioning, the original data points were broken down into K clustered sub-units to determine the centroid unit in the K*ρ partitioning scheme. The centroid unit was then used to continuously absorb and combine non-centroid units on the edges until the region demand of the integrated clustered sub-units was greater than or equal to the region's per capita efficiency.

[0016] Based on this scheme, K clustered sub-units and multiple discrete area units are finally allocated. Considering that it is difficult to accurately determine discrete area units by relying solely on the geographical isolation threshold O, this scheme also integrates global calculations and incorporates the case where multiple device locations are discrete compared to other device locations. The final area division result is K+H, where K is the number of clustered sub-units and H is the number of discrete area units.

[0017] Based on the above method, after N system partitioning simulations, the optimal solution can be selected. The optimal partitioning solution is selected based on the minimum number of discrete units and the most balanced demand of the main area.

[0018] The above division method can reasonably divide the areas of responsibility for collectors, ensuring that the workload of collectors is basically the same and ensuring fairness; on the other hand, it can cover all areas as much as possible and ensure that the number of remaining discrete units is minimized.

[0019] In a further proposed solution, step 3 specifically includes:

[0020] Step 3.1: Use the K-means clustering algorithm to initially randomly select center points for the original points on the cell distance matrix map, and assign all input parameter points to the nearest center point.

[0021] Step 3.2: Update and iterate the center point, and calculate the objective function value. When x i When it belongs to the kth class, r ik =1, otherwise r ik =0; Repeat step 4 until S no longer decreases or reaches the upper limit of the number of iterations, then stop;

[0022] Step 3.3: Regional isolation identification and population redistribution optimization. Perform regional isolation judgment on all clustered sub-units obtained in Step 3.2 iteratively, and allocate the corresponding population.

[0023] Step 3.4: Replace the center point obtained by iteration with the geographical location of the nearest device point, and assign all input parameter points to the nearest center point; break down the original K-cluster division into K*ρ clustered sub-units and H discrete region units.

[0024] In the above steps, since the straight-line distance does not match the actual traffic conditions, such as overpasses, rivers, and mountains, the cycling distance is used. In addition, considering the high cost of real-time traffic required for the algorithm to calculate distance frequently, the method of storing point distance matrix is ​​adopted for optimization: including 1. constructing a simulated matrix map in step 2 above; and 3.4 offsetting the center point to the geographical location of the nearest device point.

[0025] Due to the randomness of the initial center selection, in actual business applications, if there is significant distribution isolation among the overall locations, the number of center points (i.e., people) within mutually isolated areas may be unreasonably arranged, yet the center points cannot be iterated out. Therefore, step 3.3 is used for area isolation identification and people redistribution optimization, specifically including the following steps:

[0026] Step 1: If K <= 2, then no processing is needed for K = K; if K > 2, then perform binary clustering. If the cycling distance between the two closest points in area 1 and area 2 is greater than 0, then there is regional isolation. The total demand for area 1 (relatively small) and area 2 (relatively large) are W1 and W2 respectively. The redistribution of people in area 1 and area 2 is K = (K1, K2), where K1 = min(W1, W2) / W avg (Rounded to the nearest whole number), K2 = K - min(W1, W2) / W avg Conversely, if there is no regional isolation, then there is no need to handle K=K.

[0027] Step 2: Repeat Step 1 until (K1, K2, K3, ..., K n If no region greater than 2 or any region isolation exists, the process stops; the region to be divided is (A1, A2, A3, ..., A...). n The number of people is allocated as (K1, K2, K3, ..., K). n )

[0028] Step 3: Set the regions (A1, A2, A3, ..., A...) n If there are G regions with a population distribution greater than or equal to 1, then k-means clustering is performed to obtain the main region set G = {B1, B2, B3, ..., B}. K}, there are H elements less than 1 (H+G<=N), resulting in the discrete region set NG={B K+1 ,...,B N-G The final result is concatenated as {B1, B2, B3, ..., B}. K B K+1 ,...,B H}

[0029] In a further scheme, step 4 specifically involves: performing K-means clustering on the original data points (K = K), with the main regions defined as {B′1, B′2, B′3, ..., B′}. K}, calculate the maximum radius of each region as (B'1,R'2,...,R'). kUsing each point as a base point in turn, calculate the total required weight within a radius of 0.60 (i.e., a circular search covering approximately 1 / 3 of the area). Take the base point corresponding to the maximum value as the centroid, and take the clustered unit containing the centroid from step 4 as the centroid unit. In this scheme, after dividing the original points into K clustered sub-units, a circular search is performed covering approximately 1 / 3 of the area. When the maximum value of the equipment points is covered within the circular area, the center point of that circle is taken as the centroid.

[0030] The above scheme explains that it is difficult to accurately determine discrete areas using only the geographical isolation threshold O. Specifically, simply judging whether the distance between two adjacent device locations is greater than the geographical isolation threshold O cannot identify the overall deviation of 2-3 device locations. Therefore, it is necessary to perform a global calculation of discrete points in step 6. The specific method is as follows:

[0031] In step 1, a threshold value for the riding distance d between two adjacent device locations is defined.

[0032] Step 6 includes:

[0033] Step 6.1: Calculate and identify the discrete area units as points where the number of points within the cycling distance d threshold of all other points is less than or equal to 2.

[0034] Step 6.2: Intersect the discrete region units calculated in Step 6.1 with the discrete region units calculated in Step 5 to obtain H discrete region units.

[0035] This scheme identifies the cases where two device locations are offset from other device locations, and supplements the discrete region units that the K-means clustering algorithm failed to identify. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the k-region division method involved in the present invention.

[0037] Figure 2 The following is a schematic diagram of the clustered unit K*3 in the implementation case (the sub-region unit coefficient ρ is 3).

[0038] Figure 3 In the implementation case, the core area unit K + other area units K*2.

[0039] Figure 4 This is a schematic diagram illustrating the combination and identification of discrete points in the implementation case. Detailed Implementation

[0040] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0041] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0042] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more, unless explicitly defined otherwise.

[0043] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0044] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0045] refer to Figure 1As shown, this embodiment relates to a method for dividing recycling zones for recyclable resources by identifying clustered units and key areas, and adopts the following steps:

[0046] Step 1: Obtain the number of collectors K, the number of equipment points N, and the average daily delivery volume w of the i-th equipment. i Equipment location geographical location (longitude and latitude), geographical isolation threshold O, sub-area unit coefficient ρ, calculate per capita total area demand.

[0047] Step 2: Based on the geographical location (longitude and latitude) of the device points, obtain the distance d between each pair of points through the cycling distance API interface of the map software. ij The data is then processed and integrated into a cell distance matrix D to construct a simulated matrix map;

[0048] Step 3: The original points on the cell distance matrix map are broken down into K*ρ clustered sub-units and H discrete area units using the K-means clustering algorithm;

[0049] Step 4: Use the K-means clustering algorithm to select and divide the original points into K clustered sub-units, obtain the centroid of the K clustered sub-units, and substitute the centroid of the sub-units into the K*ρ clustered sub-units in Step 3. The clustered unit where the centroid point is located is the centroid unit.

[0050] Step 5: Using the centroid unit as the core, and following a rule from smallest to largest, match and combine other clustered units within the shortest distance O constraint to form regions. Continuously update the region with the smallest location until the demand of all regions is greater than or equal to W. avg +5000*coef [coef takes the value 0 or 1, and is defined as a parameter for whether to relax the constraints, 1 for relaxation and 0 for constraint], or when all non-centroidal units have been allocated or when they cannot be allocated due to exceeding the constraints, stop; resulting in K clustered sub-units and multiple discrete patch units;

[0051] Step 6: Calculate the discrete points globally and supplement the discrete patch units divided in Step 5 to obtain H discrete patch units;

[0052] Step 7: Repeat steps 3-6, and compare the result K+H of each area division, where K is the number of clustered sub-units and H is the number of discrete area units; take the result with the fewest remaining discrete units and the most balanced demand in the main area as the optimal area division result.

[0053] This invention employs the aforementioned technical solution, which relates to a method for dividing recyclable resource recovery zones by identifying clustered units and key areas. This method uses computer simulation for division, first setting the equipment points (recycling bin locations) within the project's scope, and the average daily delivery volume w at each equipment point.i And the number of collectors the project plans to deploy, to calculate the total demand per capita in the area.

[0054] Then, based on the GPS location of the device points in map software (such as Amap), the cycling distance between the device points is determined, thereby constructing a simulated matrix map; subsequent calculations and iterations are all performed on the simulated matrix map, thereby reducing the amount of computation.

[0055] During operation, considering that k-means has strong clustering identification capabilities but does not inherently consider demand balancing, a special design was implemented. The original K-clustering partitioning, based on a set sub-region unit coefficient ρ, broke down the original data points into K*ρ clustered sub-units and H discrete region units. Then, based on a set K-person partitioning, the original data points were broken down into K clustered sub-units to determine the centroid unit in the K*ρ partitioning scheme. The centroid unit was then used to continuously absorb and combine non-centroid units on the edges until the region demand of the integrated clustered sub-units was greater than or equal to the region's per capita efficiency.

[0056] Based on this scheme, K clustered sub-units and multiple discrete area units are finally allocated. Considering that it is difficult to accurately determine discrete area units by relying solely on the geographical isolation threshold O, this scheme also integrates global calculations and incorporates the case where multiple device locations are discrete compared to other device locations. The final area division result is K+H, where K is the number of clustered sub-units and H is the number of discrete area units.

[0057] Based on the above method, after N system partitioning simulations, the optimal solution can be selected. The optimal partitioning solution is selected based on the minimum number of discrete units and the most balanced demand of the main area.

[0058] The above division method can reasonably divide the areas of responsibility for collectors, ensuring that the workload of collectors is basically the same and ensuring fairness; on the other hand, it can cover all areas as much as possible and ensure that the number of remaining discrete units is minimized.

[0059] In a further proposed solution, step 3 specifically includes:

[0060] Step 3.1: Use the K-means clustering algorithm to initially randomly select center points for the original points on the cell distance matrix map, and assign all input parameter points to the nearest center point.

[0061] Step 3.2: Update and iterate the center point, and calculate the objective function value. When x i When it belongs to the kth class, r ik =1, otherwise r ik =0; Repeat step 4 until S no longer decreases or reaches the upper limit of the number of iterations, then stop;

[0062] Step 3.3: Regional isolation identification and population redistribution optimization. Perform regional isolation judgment on all clustered sub-units obtained in Step 3.2 iteratively, and allocate the corresponding population.

[0063] Step 3.4: Replace the center point obtained by iteration with the geographical location of the nearest device point, and assign all input parameter points to the nearest center point; break down the original K-cluster division into K*ρ clustered sub-units and H discrete region units.

[0064] In the above steps, since the straight-line distance does not match the actual traffic conditions, such as overpasses, rivers, and mountains, the cycling distance is used. In addition, considering the high cost of real-time traffic required for the algorithm to calculate distance frequently, the method of storing point distance matrix is ​​adopted for optimization: including 1. constructing a simulated matrix map in step 2 above; and 3.4 offsetting the center point to the geographical location of the nearest device point.

[0065] Due to the randomness of the initial center selection, in actual business applications, if there is significant distribution isolation among the overall locations, the number of center points (i.e., people) within mutually isolated areas may be unreasonably arranged, yet the center points cannot be iterated out. Therefore, step 3.3 is used for area isolation identification and people redistribution optimization, specifically including the following steps:

[0066] Step 1: If K <= 2, then no processing is needed for K = K; if K > 2, then perform binary clustering. If the cycling distance between the two closest points in area 1 and area 2 is greater than 0, then there is regional isolation. The total demand for area 1 (relatively small) and area 2 (relatively large) are W1 and W2 respectively. The redistribution of people in area 1 and area 2 is K = (K1, K2), where K1 = min(W1, W2) / W avg (Rounded to the nearest whole number), K2 = K - min(W1, W2) / W avg Conversely, if there is no regional isolation, then there is no need to handle K=K.

[0067] Step 2: Repeat Step 1 until (K1, K2, K3, ..., K n If no region greater than 2 or any region isolation exists, the process stops; the region to be divided is (A1, A2, A3, ..., A...). n The number of people is allocated as (K1, K2, K3, ..., K). n )

[0068] Step 3: Set the regions (A1, A2, A3, ..., A...) nIf there are G regions with a population distribution greater than or equal to 1, then k-means clustering is performed to obtain the main region set G = {B1, B2, B3, ..., B}. K}, there are H elements less than 1 (H+G<=N), resulting in the discrete region set NG={B K+1 ,...,B N-G The final result is concatenated as {B1, B2, B2, ..., B}. K B K+1 ,...,B H}

[0069] In a further scheme, step 4 specifically involves: performing K-means clustering on the original data points (K = K), with the main regions being {B′1, B′2, B′3, ..., B′}. K}, calculate the maximum radius of each region as (R′1,R′2,...,R′). K Using each point as a base point in turn, calculate the total required weight within a radius of 0.60 (i.e., a circular search covering approximately 1 / 3 of the area). Take the base point corresponding to the maximum value as the centroid, and take the clustered unit containing the centroid from step 4 as the centroid unit. In this scheme, after dividing the original points into K clustered sub-units, a circular search is performed covering approximately 1 / 3 of the area. When the maximum value of the equipment points is covered within the circular area, the center point of that circle is taken as the centroid.

[0070] The above scheme explains that it is difficult to accurately determine discrete areas using only the geographical isolation threshold O. Specifically, simply judging whether the distance between two adjacent device locations is greater than the geographical isolation threshold O cannot identify the overall deviation of 2-3 device locations. Therefore, it is necessary to perform a global calculation of discrete points in step 6. The specific method is as follows:

[0071] In step 1, a threshold value for the riding distance d between two adjacent device locations is defined.

[0072] Step 6 includes:

[0073] Step 6.1: Calculate and identify the discrete area units as points where the number of points within the cycling distance d threshold of all other points is less than or equal to 2.

[0074] Step 6.2: Intersect the discrete region units calculated in Step 6.1 with the discrete region units calculated in Step 5 to obtain H discrete region units.

[0075] This scheme identifies the cases where two device locations are offset from other device locations, and supplements the discrete region units that the K-means clustering algorithm failed to identify.

[0076] Based on the above method for identifying clustered units and key areas in the delineation of recycling zones, the following implementation case is provided for reference. Figure 2-4 As shown, Figure 1 It is a 3*3 (K=3, 3 times scattered) clustered unit identification. In the first step, the '+' area [9,17,13,45,2,8,32,6] is identified as an isolated area with a k value of 1. In the second step, the 'equilateral triangle' area [7,39,25] is identified as an isolated area. During redistribution, due to the low total demand, the k value is rounded to 0, making it a discrete area not included in the arrangement. After the third step of calculation, it is found that there are no isolated areas. The remaining k value is allocated to 8, and finally 9 clustered area units to be combined are formed. There is also a discrete area unit [7,39,25]. At this time, the arrangement in the upper right corner has no center of gravity.

[0077] Figure 2 It temporarily performs 3*1 centroid area unit identification, and identifies [9,17,13,45,2,8,32,6], [33,20,18,21,42], and [26,28,38,31,0] as centroid area units, and the remaining areas as other area units;

[0078] Figure 3 It involves area combination and discrete point identification, using the core area as the core and sorting according to the total demand from low to high. Each time, the smallest core area is combined with other area units that meet the distance limit. The core area is continuously updated and reordered until no more units meet the limit or all other units are allocated. After combination, discrete points are identified that are less than two points away from all other points within a cycling distance threshold O, ultimately forming a pattern like... Figure 3 The final results of the region division are as follows: '+' region [9,17,13,45,2,8,32,6], 'pentagon' region [1,10,18,20,21,33,34,37,42,44,4,5,30,46], 'prism' region [11,12,19,22,27,29,35,41,43,0,15,24,26,28,31,36,38,40], discrete 'equilateral triangle' region [7,39,25], and the remaining discrete points [9,17,3,14]. End.

[0079] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0080] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention without departing from the principles and spirit of the present invention.

Claims

1. A method for dividing recycling zones into areas that identify clustered units and key regions, characterized in that: The following steps are used: Step 1: Obtain the number of collectors K, the number of equipment points N, and the average daily delivery volume w of the i-th equipment. i Equipment location, geographical isolation threshold O, sub-area unit coefficient ρ, calculate the per capita total demand in the area = ; Step 2: Based on the geographical location of the device, obtain the distance d between each pair of points using the cycling distance API interface of the map software. ij And process and integrate them into a cell distance matrix D; Step 3: The original points on the cell distance matrix map are broken down into K*ρ clustered sub-units and H discrete area units using the K-means clustering algorithm; Step 3 specifically includes: Step 3.1: Use the K-means clustering algorithm to initially randomly select center points for the original points on the cell distance matrix map, and assign all input parameter points to the nearest center point; Step 3.2: Update and iterate the center point, and calculate the objective function value S= When x i When it belongs to the Kth class, r iK =1, otherwise r iK =0; Repeat this step until S no longer decreases or reaches the upper limit of the number of iterations, then stop; Step 3.3: Regional isolation identification and population redistribution optimization. Perform regional isolation judgment on all clustered sub-units obtained in Step 3.2 iteratively, and allocate the corresponding population. Step 3.4: Replace the center point obtained through iteration with the geographical location of the nearest device point, and assign all input parameter points to the nearest center point; break down the original K-cluster division into K*ρ clustered sub-units and H discrete region units; Step 3.3 includes the following steps: Step 1: If K <= 2, then no processing is needed for K = K; if K > 2, then perform binary clustering. If the cycling distance between the two closest points in area 1 and area 2 is greater than the geographical isolation threshold O, then regional isolation exists. The total demand in area 1 and area 2 are W1 and W2 respectively. The redistribution of people in area 1 and area 2 is K = (K1, K2), where K1 = min(W1, W2) / W avg Round K1 to the nearest integer, and K2 = K - min(W1, W2) / W avg Conversely, if there is no regional isolation, then there is no need to handle K=K; Step 2: Repeat Step 1 until (K1, K2, K3, ..., K n If no region greater than 2 or any region isolation exists, the process stops; the region to be divided is (A1, A2, A3, ..., A...). n The number of people is allocated as (K1, K2, K3, ..., K). n ); Step 3: Set the regions (A1, A2, A3, ..., A...) n If there are G regions with a population distribution greater than or equal to 1 in the given data, then K-means clustering will yield the main region set G = {B1, B2, B3, ..., B}. K }, there are H elements less than 1, H+G<=N, so the discrete region set NG={B K+1 B N-G The final result is concatenated as {B1, B2, B3, ..., B}. K B K+1 B H }; Step 4: Use the K-means clustering algorithm to divide the original points into K clustered sub-units, obtain the centroids of these K clustered sub-units, and substitute these centroids into the K*ρ clustered sub-units from Step 3. The clustered unit containing the centroid is the centroid unit. Specifically, Step 4 involves performing K-means clustering on the original points (K = K), with the main region being {B...} ′ 1, B ′ 2, B ′ 3, ..., B ′ K }, calculate the maximum radius of each region as (R ′ 1, R ′ 2, ..., R ′ K Taking each point as a base, calculate the total required weight within a radius of 0.60, take the base point corresponding to the maximum value as the centroid, and take the clustering unit where the centroid in step 4 is located as the centroid unit; Step 5: Using the centroid unit as the core, and following a rule from smallest to largest, match and combine other clustered units within the geographical isolation threshold O to form areas. Continuously update the area with the smallest location until the demand for all areas is greater than or equal to W. avg +5000*coef, where coef takes the value 0 or 1 and is defined as a parameter for whether to relax the constraints. 1 means relaxation and 0 means constraint. The process stops when non-centroidal units are all assigned or when they cannot be assigned due to exceeding the constraints. This results in K clustered sub-units and multiple discrete patch units. Step 6: Calculate the discrete points globally and supplement the discrete patch units divided in Step 5 to obtain H discrete patch units; Step 7: Repeat steps 3-6, and compare the result K+H of each area division, where K is the number of clustered sub-units and H is the number of discrete area units; take the result with the fewest remaining discrete units and the most balanced demand in the main area as the optimal area division result.

2. The method for dividing recyclable resource recovery zones according to claim 1, characterized in that: In step 1, a threshold value for the riding distance d between two adjacent device locations is defined. Step 6 includes: Step 6.1: Calculate and identify the discrete area units as points where the number of points within the cycling distance d threshold of all other points is less than or equal to 2. Step 6.2: Intersect the discrete region units calculated in Step 6.1 with the discrete region units calculated in Step 5 to obtain H discrete region units.

Images