Agricultural and forestry waste recycling method and system based on big data

Abstract

The application provides a kind of based on big data's forestry waste recycling method and system, obtains forestry waste generation point and candidate recycling point information;The planned area is gridded, determines the weight of building point, and accordingly generates initial population, individual coding includes point layout and initial scheduling path;Generate offspring by rotating center crossover operation: select a point as the center of circle, exchange the point and path gene segment in the two concentric ring regions of parent generation, and repair the path after crossover;Elite retention strategy is used to select individuals to form a new generation population;If the average fitness of the population stagnates continuously for several generations, perform variable neighborhood search on the current optimal individual: close one or two points, select the point closest to the center of gravity in the service point to open a new point, and optimize the affected path by two-point exchange and node insertion, replace the new individual back to the population;Repeat iteration until the maximum number of iterations is reached, output the recycling point layout and vehicle scheduling scheme corresponding to the optimal individual.

Description

Technical Field

[0001] This application belongs to the field of recycling and processing, and in particular relates to a method and system for recycling and processing agricultural and forestry waste based on big data. Background Technology

[0002] The recycling logistics system hinges on the layout of recycling points and the scheduling of transport vehicles, both of which jointly determine the total cost and operational efficiency of the entire recycling network. In operations research, this problem is categorized as the Location Path Problem (LRP), an NP-hard combinatorial optimization problem. Planning methods, based on human experience or simple models, struggle to find globally optimal or near-optimal solutions under realistic constraints, hindering the industrialization of agricultural and forestry waste. Exact algorithms, such as branch and bound and cutting plane methods, can theoretically find the optimal solution, but their computational complexity increases exponentially with the problem size. For large-scale agricultural and forestry waste recycling network planning, which often involves hundreds or thousands of waste generation points, the computation time becomes unacceptably long. Therefore, heuristic algorithms, especially metaheuristic algorithms such as genetic algorithms, simulated annealing, tabu search, and ant colony optimization, can find satisfactory solutions within a reasonable timeframe. However, when using a univariate heuristic genetic algorithm to solve this problem, the algorithm is prone to premature convergence, with population diversity decreasing in the later stages of the search, leading to getting trapped in local optima and unable to escape. While standard crossover and mutation operators can perform global searches, their ability to deeply mine high-quality solutions is insufficient. Furthermore, for encodings that simultaneously contain layout and path information, general genetic operators can disrupt the optimal structure of the solution, resulting in low search efficiency. Therefore, there is an urgent need for an improved optimization method that integrates global and local searches and avoids premature convergence, to obtain a better layout and scheduling scheme for agricultural and forestry waste recycling points. Summary of the Invention

[0003] This invention proposes a big data-based method for the recycling and processing of agricultural and forestry waste, addressing the lack of an improved optimization method in existing technologies that integrates global and local search and avoids premature convergence. The method includes:

[0004] Obtain the location and quantity information of agricultural and forestry waste generation points and candidate recycling point location information within the area to be planned; divide the planning area into grids, determine the point construction weight based on the total amount of waste generated in each grid, and generate an initial population based on the point construction weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path.

[0005] Iterative optimization is performed using total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population.

[0006] During the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, a variable neighborhood search is initiated for local optimization of the current best individual, and the optimized new individual is replaced back into the population.

[0007] Repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

[0008] This invention also proposes a big data-based agricultural and forestry waste recycling and processing system, comprising the following modules:

[0009] The generation module is used to obtain the location and quantity information of agricultural and forestry waste generation points and the location information of candidate recycling points in the area to be planned; divide the planning area into grids, determine the construction point weight according to the total amount of waste generated in each grid, and generate an initial population based on the construction point weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path.

[0010] The selection module is used for iterative optimization with total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population.

[0011] The execution module is used to, during the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, initiate a variable neighborhood search for local optimization of the current best individual, and replace the optimized new individual back into the population;

[0012] The output module is used to repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

[0013] This invention integrates the modeling of recycling point layout and vehicle scheduling problems, and uses the point construction weight based on the total amount of waste generated to guide the generation of the initial population, resulting in a high-quality initial solution that closely matches actual needs. In iterative optimization, concentric ring crossover operations are employed to preserve superior gene fragments from the parent generation, enhancing the algorithm's global search capability. To address the potential for the algorithm to get stuck in local optima, a strategy combining stagnation detection and variable neighborhood search is devised. By performing deep local optimization on the current optimal solution, the algorithm escapes local extrema, improving its optimization accuracy and achieving the lowest-cost integrated solution for recycling point layout and vehicle scheduling. This solution is rationally laid out, efficiently scheduled, and possesses both scientific merit and practicality. Attached Figure Description

[0014] Figure 1 A flowchart of the first embodiment;

[0015] Figure 2 This is a schematic diagram of the initial data distribution;

[0016] Figure 3 This is a diagram illustrating a two-point exchange operation. Detailed Implementation

[0017] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0018] In the first embodiment, the present invention proposes a method for recycling and processing agricultural and forestry waste based on big data, see [link to relevant documentation]. Figure 1 ,include:

[0019] S1, obtain the location and quantity information of agricultural and forestry waste generation points and candidate recycling point location information in the area to be planned; divide the planning area into grids, determine the construction point weight according to the total amount of waste generated in each grid, and generate an initial population based on the construction point weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path.

[0020] Using Geographic Information System (GIS) software, import the administrative map of the area to be planned. Based on field survey data, mark each farm or village on the map as a waste generation point, and add latitude and longitude coordinates and average annual straw production tons as attribute fields. Simultaneously, mark existing storage centers, transportation hubs, or idle land locations within the area as candidate recycling sites on the map. (See [link to relevant documentation]). Figure 2 And record the latitude and longitude coordinates.

[0021] In the geographic information system software, a grid is created on the map of the planning area with a size of 1 km × 1 km. The output tons of all waste generation points within each grid cell are counted and accumulated, and this tonnage is the grid's point weight. The population size is set to 100 individuals, and each individual consists of two parts of coding. The first part is the recycling point layout code, which is an integer list representing the ID number of the selected candidate recycling points. The second part is the vehicle scheduling path code, which is a set of multiple sublists. Each sublist represents a vehicle path and contains the recycling point ID and the sequence of service generation point IDs. An initial layout is generated using a roulette wheel selection method, where the candidate recycling point with a larger grid point weight has a higher probability of being selected. After generating the layout scheme, each waste generation point is assigned to the nearest selected recycling point. For the set of generation points assigned to each recycling point, an initial access path is generated using the nearest neighbor algorithm with that recycling point as the start and end point.

[0022] In an optional embodiment, dividing the planning area into grids and determining the site weights based on the total amount of waste generated in each grid includes:

[0023] The planned area is divided into N grids of preset size;

[0024] Calculate the total amount of waste within each grid;

[0025] The point weight of the i-th grid is the ratio of the total amount of waste in the i-th grid to the total amount of waste in all grids of the planning area.

[0026] Suppose a 10 km × 10 km planning area is divided into 100 grids of 1 km × 1 km each, i.e., N = 100. Through field surveys or historical data, the total daily waste generated by all waste-generating points within each grid is calculated. For example, if grid number 5 contains three waste-generating points with daily generation amounts of 1.2 tons, 0.8 tons, and 1.5 tons respectively, then the total waste generated by this grid is... The total is 3.5 tons. This process is repeated for all 100 grids to obtain the total for each grid. , , , .

[0027] After calculating the total waste volume for all grids, the total waste volume for the entire planning area is calculated, for example, 200 tons. The point weight for each grid is then calculated using a formula. For grid number 5, the point weight is... The sum of the weights of all grid cells is 1, forming a probability distribution. During the population initialization phase of the genetic algorithm, a roulette wheel selection is performed based on these weights to determine the initial layout of recycling points. Grid cells with larger weights, i.e., areas with more concentrated waste generation, have a higher probability of being selected as recycling points, thus making the initial solution closer to high-quality regions and improving the algorithm's convergence efficiency.

[0028] S2, iterative optimization is performed using total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population.

[0029] The total cost equals the sum of the fixed costs of constructing and operating recycling points and the total mileage cost of vehicle transportation. The fitness value is the reciprocal of the total cost. Two parent individuals, A and B, are selected from the population. A recycling point is randomly selected from the layout scheme of parent A, for example, a point with ID 10. Using the coordinates of this point as the center, a ring-shaped area is randomly generated with an inner radius of 5 kilometers and an outer radius of 15 kilometers. All recycling points and waste generation points in parents A and B whose coordinates lie within this ring-shaped area are identified, forming their respective gene segments. When generating offspring A, the parent A's gene segments are retained. Genes outside the circular region are copied from the circular region gene segment of parent generation B; offspring A is repaired, and it is checked whether any waste generation points have been assigned to recycling points that no longer exist. If so, they are reassigned to the recycling points in the nearest new layout. For all vehicle paths that have changed due to the exchange, the shortest travel route is recalculated; offspring B is generated using the same method; 100 parent individuals and 100 cross-generated offspring individuals are merged, the total cost of each individual is calculated, and they are sorted from low to high total cost. The top 100 individuals with the lowest cost are selected to form the new generation population.

[0030] In an optional embodiment, using total cost as the fitness function includes:

[0031] The total cost Construction and operation costs of recycling outlets and vehicle transportation costs It consists of two parts, and the calculation formula is:

[0032]

[0033] Where S is the set of enabled recycling points, To recover the fixed construction and operation costs of point j, K is the set of all vehicle dispatch routes. Let k be the distance traveled along path k. Cost per unit distance;

[0034] The reciprocal of the total cost is used as the individual's fitness value.

[0035] Cost accounting is performed on a specific site selection and route planning scheme, i.e., an individual in the genetic algorithm. The model structure of this scheme consists of two parts: first, the selected set of recycling points; and second, the set of vehicle scheduling routes from these points to the waste generation points they serve. For example, a scheme might decide to open three recycling points numbered 1, 2, and 3, with annual fixed construction and operation costs... The costs are 100,000 yuan, 120,000 yuan, and 90,000 yuan respectively. Therefore, the total fixed cost is 310,000 yuan. The scheduling path satisfies vehicle capacity and maximum travel constraints.

[0036] Assuming that 5 vehicle recycling routes are planned for the three recycling points, the total driving distance of the routes is... The distances are 50 km, 45 km, 60 km, 55 km, and 40 km respectively. What is the transportation cost per unit distance? Setting the cost at 10 yuan per kilometer, the total transportation cost is 2500 yuan. Therefore, the total cost of this scheme is 312,500 yuan. To unify the optimization objective into a maximization problem suitable for the selection mechanism of the genetic algorithm, the reciprocal of the total cost is taken as the fitness value of the individual. The lower the cost of the scheme, the higher the fitness value, and the greater the probability of it being retained and passed on to the next generation in the genetic operation.

[0037] In an optional embodiment, the crossover operation includes:

[0038] Randomly select an open recycling point from one of the parent individuals to be crossed as the center;

[0039] Using the preset inner and outer radii, concentric ring regions are defined on the codes of the two parent individuals;

[0040] The pathway gene segments formed by the recycling network point genes and the waste generation points served by the two parent individuals located within the concentric ring area are exchanged.

[0041] Suppose a crossover operation is performed between parent individuals A and B. A random point is selected from parent A's active recycling points, for example, point A5 with geographical coordinates (30, 40), as the center of the crossover operation. The inner circle radius is preset to 5 kilometers, and the outer circle radius to 15 kilometers, thus defining a concentric ring region on the entire planning area map. This region represents a specific geographical neighborhood.

[0042] Scanning the coding of parent A revealed that the activated recycling point A8 and the waste generation points P1, P2, and P3 served by A8 all fall within this circular area. This portion of the gene coding collectively constitutes a gene segment of parent A. Similarly, scanning the coding of parent B revealed that the activated recycling points B3 and B9 and the waste generation points P4, P5, and P6 served by them are also located in this area, constituting a gene segment of parent B. The gene coding is a complex structure that simultaneously contains location information and associated path information.

[0043] The gene segments identified in parent A (network point A8 and path P1-P2-P3) are exchanged with the gene segments identified in parent B (network points B3, B9 and path P4-P5-P6). The resulting offspring, individual one, has a basic structure derived from parent A, but the portion within the circular region is replaced with the corresponding gene segments from parent B. Offspring individual two is generated similarly. This geographically neighborhood-based crossover method helps maintain and recombine excellent local regional layout structures, thereby accelerating the algorithm's exploration towards a better solution.

[0044] In an optional embodiment, the repair of broken and reconnected cross-path segments resulting from the swap includes:

[0045] For offspring individuals after crossover, identify the points of origin for unclaimed waste generated due to gene exchange;

[0046] The unclaimed waste generation points are reassigned to the nearest recycling point that is active in the offspring individuals;

[0047] The reassigned waste generation points are inserted into the existing scheduling paths of the newly assigned recycling points, based on the principle of minimizing path increments, to form the repaired new paths.

[0048] After the crossover operation is completed, a offspring individual is generated. Suppose that in parent A, there is a waste generation point P10 that was originally served by recycling point A8. However, during crossover, a gene segment from parent B replaces the gene segment containing A8. Therefore, in the new offspring individual, recycling point A8 no longer exists, causing P10 to lose its serving recycling point and become an unclaimed waste generation point. The repair operation involves traversing all waste generation points of the offspring individual and finding all points similar to P10. Optionally, after the swap, duplicate waste generation points in the offspring are deleted before processing the unclaimed points.

[0049] For each unclaimed waste generation point P10, calculate its Euclidean distance to all active recycling points in the current offspring. For example, if there are active recycling points A1, A2, and B3 in the offspring, the calculated distances from P10 to these three points are 12 km, 15 km, and 8 km, respectively. Select the nearest recycling point B3 as the new designated service point for P10.

[0050] After determining that P10 belongs to point B3, it is inserted into the existing vehicle scheduling path of B3. Assume the original path of B3 is "B3-P4-P5-B3". To minimize the impact on the total length of the original path, it is necessary to try inserting P10 into all possible positions, such as "B3-P10-P4-P5-B3", "B3-P4-P10-P5-B3", and "B3-P4-P5-P10-B3". The increase in the total path length caused by each of these three insertion methods is calculated, and the scheme with the smallest increase is selected. For example, inserting it between P4 and P5 results in the shortest increase in path mileage, so the repaired new path is "B3-P4-P10-P5-B3". Through this series of repair steps, it is ensured that the offspring individuals generated after the crossover operation are all complete solutions.

[0051] S3, During the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, a variable neighborhood search is initiated for local optimization of the current best individual, and the optimized new individual is replaced back into the population.

[0052] In one embodiment, the variable neighborhood search includes: randomly closing one or two recycling points, selecting the point closest to the centroid from the waste-generating points served by the closed recycling points as the new recycling point to open, and performing two-point swaps and node insertions on all affected scheduling paths.

[0053] The stagnation generation is set to 20 generations. If the average total cost of the population decreases by less than 0.01% over 20 consecutive generations, a variable neighborhood search is triggered. The individual with the lowest total cost in the current population is selected, and a recycling point in the layout scheme is randomly chosen, for example, point ID 8, and this point is closed. All waste generation points previously served by point 8 are identified, their geometric center points are calculated, and the distance from each generation point to this center point is calculated. The closest generation point, for example, P5, is selected, and the candidate recycling point closest to P5 is set to open, replacing the closed point 8. The generation points previously served by point 8 are reassigned to the recycling points closest to them in the new layout. For all vehicles whose paths have changed, a two-point exchange optimization is performed (see [link to relevant documentation]). Figure 3 This involves attempting to shorten the path by swapping the access order of any two generating points in the path, and performing node insertion optimization, which involves trying to remove a generating point from the current path and insert it into every possible position in another path. If this reduces the total transportation cost, the move is retained. The locally optimized new individual is compared with the individual with the highest cost in the population; if the new individual has a lower cost, the worst individual is replaced.

[0054] In an optional embodiment, the detection that the average fitness of the population has stagnated for a specified number of generations specifically means: if for G consecutive generations, the average fitness value of the population... The rate of change is lower than the preset threshold If so, it is determined to be stagnant.

[0055] In an optional embodiment, selecting the point closest to the center of gravity from the waste-generating points served by the closed recycling points as the new recycling point to be opened includes:

[0056] Identify the set of waste generation points that become unserviced due to the closure of recycling points, and calculate the geographic centroid coordinates of the set;

[0057] Among the candidate recycling points, the candidate point with the closest Euclidean distance to the geographical centroid is selected as the newly opened recycling point.

[0058] When performing a perturbation operation to escape local optima, a recycling point with low utility in the current solution, such as point C, is selected for closure. After closure, all waste generation points previously served by point C become unserviced. Assume the set of points is P (unserviced), containing four points: P1, P2, P3, and P4, with coordinates P1 (10, 20), P2 (15, 25), P3 (12, 18), and P4 (18, 22), respectively. Calculate the geocentric coordinates of this point set, where the x-coordinate is the average of all x-coordinates (13.75) and the y-coordinate is the average of all y-coordinates (21.25).

[0059] Having a list of all potential candidate recycling points, we need to identify those not currently active in the current scheme. Assume there are three available candidate points: Candidate 1 (coordinates (14, 21)); Candidate 2 (coordinates (19, 30)); and Candidate 3 (coordinates (8, 15)). Calculate the Euclidean distance between the geographic centroid obtained in the previous step (coordinates (13.75), (21.25)) and each candidate recycling point. The distance to Candidate 1 is approximately 0.35, to Candidate 2 is approximately 10.16, and to Candidate 3 is approximately 8.54. Comparing these distances, Candidate 1 is found to be closest to the centroid. Therefore, Candidate 1 is selected as the new recycling point to be activated. The strategy is to re-establish a new service point at the center of an area where services have been withdrawn, covering the newly unserviced waste generation points, thereby optimizing the layout.

[0060] In an optional embodiment, performing two-point swapping and node insertion on all affected scheduling paths includes:

[0061] Waste generation points affected by changes in the layout of recycling points will be reassigned to the nearest active recycling point to determine the set of affected scheduling paths.

[0062] For each affected scheduling path, iteratively perform two-point swap operations and node insertion operations to optimize it until the total path length can no longer be shortened.

[0063] The two-point swap is to randomly swap the positions of two waste generation points in the path, and the node insertion is to attempt to insert a waste generation point in the path into all other possible positions.

[0064] When the layout of recycling points changes, for example, when point C is closed and point D is opened, all waste generation points previously served by point C need to be reassigned. The new distance to all currently open points is calculated for each point, and it is assigned to the nearest point. For example, point P1, originally belonging to point C, is assigned to point A, while points P2 and P3 are assigned to point B. The vehicle dispatching routes of points A and B change due to the addition of the new service point, forming the set of affected dispatching routes.

[0065] For each affected path in the set, such as the new path to point A, a local search optimization is performed. This optimization process includes two core operations: point swapping and node insertion. The point swapping operation involves randomly selecting two waste-generating points in the path sequence and exchanging their positions. For example, if the original path is "A-P10-P11-P1-A", swapping P10 and P1 results in "A-P1-P11-P10-A", and the total length of the new path is calculated to determine if it has been shortened.

[0066] The node insertion operation involves removing a point, such as P1, from its current position in the path and attempting to insert it into all other possible positions in the path sequence. The path length for each insertion scheme is calculated, and the scheme that minimizes the path length is retained. This process of alternating between point swaps and node insertions continuously updates the path. This iterative process continues until performing any further point swaps or node insertions on the path fails to shorten the total length. After optimizing the path for point A, the same optimization process is performed on the path for point B, ensuring that all paths affected by layout changes reach a local optimum.

[0067] S4. Repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

[0068] The maximum number of iterations is set to 500. When the number of iterations reaches 500, the algorithm stops and selects the individual with the lowest total cost from the population. The decoded information of this individual is output, including: a list of identified recycling point IDs and their geographical coordinates; the specific driving route of each transport vehicle, in the format of starting from recycling point A -> generating point 3 -> generating point 15 -> generating point 9 -> returning to recycling point A; and the lowest total cost value under this scheme.

[0069] In the second embodiment, the present invention also proposes a big data-based agricultural and forestry waste recycling and processing system, comprising the following modules:

[0070] The generation module is used to obtain the location and quantity information of agricultural and forestry waste generation points and the location information of candidate recycling points in the area to be planned; divide the planning area into grids, determine the construction point weight according to the total amount of waste generated in each grid, and generate an initial population based on the construction point weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path.

[0071] The selection module is used for iterative optimization with total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population.

[0072] The execution module is used to, during the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, initiate a variable neighborhood search for local optimization of the current best individual, and replace the optimized new individual back into the population;

[0073] The output module is used to repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

[0074] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

Claims

1. A method for recycling and processing agricultural and forestry waste based on big data, characterized in that, Includes the following steps: Obtain the location and quantity information of agricultural and forestry waste generation points and candidate recycling point location information within the area to be planned; divide the planning area into grids, determine the point construction weight based on the total amount of waste generated in each grid, and generate an initial population based on the point construction weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path. Iterative optimization is performed using total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population. During the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, a variable neighborhood search is initiated for local optimization of the current best individual, and the optimized new individual is replaced back into the population. Repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

2. The method according to claim 1, characterized in that, The step of dividing the planned area into grids and determining the weight of each construction point based on the total amount of waste generated in each grid includes: The planned area is divided into N grids of preset size; Calculate the total amount of waste within each grid; The point weight of the i-th grid is the ratio of the total amount of waste in the i-th grid to the total amount of waste in all grids of the planning area.

3. The method according to claim 1, characterized in that, The crossover operation includes: Randomly select an open recycling point from one of the parent individuals to be crossed as the center; Using the preset inner and outer radii, concentric ring regions are defined on the codes of the two parent individuals; The pathway gene segments formed by the recycling network point genes and the waste generation points served by the two parent individuals located within the concentric ring area are exchanged.

4. The method according to claim 1, characterized in that, The repair of broken and reconnected cross-path segments resulting from the exchange includes: For offspring individuals after crossover, identify the points of origin for unclaimed waste generated due to gene exchange; The unclaimed waste generation points are reassigned to the nearest recycling point that is active in the offspring individuals; The reassigned waste generation points are inserted into the existing scheduling paths of the newly assigned recycling points, based on the principle of minimizing path increments, to form the repaired new paths.

5. The method according to claim 1, characterized in that, The variable neighborhood search includes: randomly closing one or two recycling points, selecting the point closest to the centroid from the waste generation points served by the closed recycling points as the new recycling point, and performing two-point swaps and node insertions on all affected scheduling paths.

6. The method according to claim 5, characterized in that, The step of selecting the point closest to the center of gravity from the waste generation points served by the closed recycling points as the new recycling point to be opened includes: Identify the set of waste generation points that become unserviced due to the closure of recycling points, and calculate the geographic centroid coordinates of the set; Among the candidate recycling points, the candidate point with the closest Euclidean distance to the geographical centroid is selected as the newly opened recycling point.

7. The method according to claim 5, characterized in that, The process of performing two-point swaps and node insertions on all affected scheduling paths includes: Waste generation points affected by changes in the layout of recycling points will be reassigned to the nearest active recycling point to determine the set of affected scheduling paths. For each affected scheduling path, iteratively perform two-point swap operations and node insertion operations to optimize it until the total path length can no longer be shortened. The two-point swap is to randomly swap the positions of two waste generation points in the path, and the node insertion is to attempt to insert a waste generation point in the path into all other possible positions.

8. A big data-based agricultural and forestry waste recycling and processing system, characterized in that, Includes the following modules: The generation module is used to obtain the location and quantity information of agricultural and forestry waste generation points and the location information of candidate recycling points in the area to be planned; divide the planning area into grids, determine the construction point weight according to the total amount of waste generated in each grid, and generate an initial population based on the construction point weight. The code of each individual in the population includes the recycling point layout scheme and the initial vehicle scheduling path. The selection module is used for iterative optimization with total cost as the fitness function. The iterative optimization includes: generating offspring individuals through a crossover operation, wherein the crossover operation involves randomly selecting a recycling point as the rotation center, generating two concentric circles with the center as the center, exchanging the point gene segments and path gene segments located within the concentric circles of the two parent individuals, and repairing the broken and reconnected crossover path segments generated after the exchange; and selecting individuals from the parent and offspring generations according to the fitness value using an elite retention strategy to form a new generation population. The execution module is used to, during the iteration process, if the average fitness of the population is detected to have stagnated for a specified number of generations, initiate a variable neighborhood search for local optimization of the current best individual, and replace the optimized new individual back into the population; The output module is used to repeat the iterative process until the preset maximum number of iterations is reached, and output the recycling point layout and scheduling scheme corresponding to the individual with the best fitness in the population.

9. The system according to claim 8, characterized in that, The step of dividing the planned area into grids and determining the weight of each construction point based on the total amount of waste generated in each grid includes: The planned area is divided into N grids of preset size; Calculate the total amount of waste within each grid; The point weight of the i-th grid is the ratio of the total amount of waste in the i-th grid to the total amount of waste in all grids of the planning area.

10. The system according to claim 8, characterized in that, The crossover operation includes: Randomly select an open recycling point from one of the parent individuals to be crossed as the center; Using the preset inner and outer radii, concentric ring regions are defined on the codes of the two parent individuals; The pathway gene segments formed by the recycling network point genes and the waste generation points served by the two parent individuals located within the concentric ring area are exchanged.

Images