An agricultural resource scheduling method and system based on smart villages

Abstract

The present application relates to resource scheduling technical field, especially to a kind of agricultural resource scheduling method and system based on wisdom village, method includes steps: the road network distance of order destination to starting point and distribution path are acquired;By analyzing the distribution characteristics of road network distance and the similarity of distribution path, determine initial clustering center in two-stage screening mode, obtain multiple distribution subareas based on the initial clustering center clustering;Integer linear programming model is constructed considering vehicle loading utilization rate and estimated transportation cost, for each distribution subarea dynamically matching optimal vehicle combination.The present application solves the randomness and unreasonableness of initial clustering center selection of traditional clustering algorithm, ensures that distribution subarea result is optimal in physical distance and driving direction;For each distribution subarea dynamically matching optimal vehicle combination, reduce the additional cost generated due to transport capacity waste or improper vehicle selection.

Description

Technical Field

[0001] This invention relates to the field of resource scheduling technology, and in particular to an agricultural resource scheduling method and system based on smart villages. Background Technology

[0002] In the logistics and distribution of agricultural resources such as fertilizers and seeds, companies generally face the challenge of handling a large number of scattered and time-sensitive orders during peak sales seasons. The common strategy in the industry is to first divide the orders into zones and then plan the routes for each zone.

[0003] Currently, the common technique for implementing order partitioning is to use clustering algorithms such as K-Means. These algorithms iterate to minimize the sum of squared distances from each order's destination to the centroid of its partition, but they rely entirely on Euclidean distance to measure the proximity of orders.

[0004] However, this traditional scheduling method has significant drawbacks in practical applications: Firstly, clustering algorithms such as K-Means exhibit high randomness in selecting initial cluster centers, leading to slow convergence and a high likelihood of getting trapped in local optima, resulting in unstable and inconsistent partitioning results. Secondly, and more critically, Euclidean distance cannot perceive the real-world road network topology. For example, two order destinations separated by rivers or mountains may be close in a straight line, but far apart in terms of road network distance. Traditional clustering algorithms frequently misclassify such orders into the same partition, resulting in partitioning schemes that are completely infeasible or extremely inefficient in reality. Drivers are forced to take numerous unnecessary detours, significantly increasing transportation costs and defeating the purpose of partitioning. Therefore, overcoming the randomness in the selection of initial cluster centers and the limitations in distance metrics of traditional clustering algorithms is a pressing technical challenge in the field of resource scheduling. Summary of the Invention

[0005] To address the technical problem that the randomness in the initial cluster center selection of the aforementioned clustering algorithm easily leads to unstable partitioning results, affecting the rationality of delivery planning, and thus reducing overall resource scheduling efficiency and increasing transportation costs, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an agricultural resource scheduling method based on smart villages, the method comprising the steps of:

[0007] The system acquires road network distances and delivery routes from multiple order destinations to a preset origin; determines candidate cluster centers based on the distribution characteristics of the road network distances, and filters the candidate cluster centers based on the similarity of the delivery routes to obtain initial cluster centers; combines the initial cluster centers and road network distances to cluster the multiple order destinations to obtain multiple delivery zones; based on the total order weight and the furthest road network distance within each delivery zone, and combined with preset vehicle cost parameters, matches the optimal vehicle combination between vehicle loading utilization and estimated transportation costs for each delivery zone; and performs vehicle route planning for the optimal vehicle combination within each delivery zone to obtain the driving routes of each vehicle, thereby completing the scheduling of agricultural resources.

[0008] This invention first identifies order clusters based on the distribution characteristics of road network distances through statistical analysis. Then, it filters based on the similarity of delivery routes, overcoming the randomness and irrationality of the initial cluster center selection in traditional clustering algorithms. This ensures that the partitioning results are optimal in both physical distance and driving direction, improving the feasibility of partitioning. Furthermore, by comprehensively considering vehicle loading utilization and estimated transportation costs, this invention dynamically matches the optimal vehicle combination for each delivery partition, reducing additional costs caused by wasted capacity or improper vehicle selection. Ultimately, it achieves a dual optimization of overall scheduling efficiency and cost control.

[0009] Preferably, determining candidate cluster centers based on the distribution characteristics of the road network distances includes: statistically analyzing the road network distances from all order destinations to the preset starting point and constructing a distance histogram; determining the peak intervals in the distance histogram as clustering regions, and selecting an order destination as a candidate cluster center within each clustering region.

[0010] This invention provides a data-driven, automated method to identify order clusters by constructing a distance histogram and identifying peak intervals. This method avoids the uncertainty and local optima problems caused by randomness, making the selected candidate cluster centers highly representative and laying the foundation for generating high-quality delivery partitions.

[0011] Preferably, selecting an order destination as a candidate cluster center within each cluster region includes: calculating the average of the geographical coordinates of all order destinations within the cluster region to obtain the geographical center point; and selecting the order destination within the cluster region that is closest to the geographical center point as the candidate cluster center.

[0012] This invention calculates the geographic center of all orders within a region and selects the actual order closest to that center as the candidate cluster center. This ensures that the selected candidate cluster center is the most geographically representative point within its cluster region, thereby avoiding clustering bias that may be caused by selecting edge points and improving the quality of candidate cluster centers.

[0013] Preferably, the step of filtering candidate cluster centers based on the similarity of the delivery routes to obtain initial cluster centers includes: for each candidate cluster center, recording the set of order destinations within its corresponding cluster area as a candidate partition, calculating the average path similarity within each candidate partition; and recording candidate cluster centers with an average path similarity greater than a preset similarity threshold as initial cluster centers.

[0014] This invention addresses the shortcomings of existing technologies that rely on distance clustering to distinguish between order destinations that are the same distance away but in opposite directions. Traditional methods may incorrectly cluster two order destinations that are equally far from the warehouse but in opposite directions, but this method identifies this discrepancy due to their low path similarity. This ensures that the final delivery zones are not only compact in distance but also highly consistent in delivery routes, avoiding unnecessary detours and making the delivery zone results match the actual logistics delivery goals.

[0015] Preferably, the calculation of the average path similarity within each candidate partition includes: using a dynamic time warping algorithm to calculate the similarity between any two delivery paths within the candidate partition.

[0016] Preferably, obtaining the optimal vehicle combination includes: for each delivery zone, calculating the difference between the total rated load of the vehicle combination and the total order weight, and recording it as the first optimization objective; based on preset vehicle cost and variable cost parameters, combined with the farthest road network distance of the delivery zone, calculating the unit transportation cost of each available vehicle, and calculating the sum of the unit transportation costs of all vehicles in the vehicle combination, and recording it as the second optimization objective; constructing an objective function with optimization indicators based on the first and second optimization objectives; and solving the objective function using an integer linear programming algorithm to obtain the optimal vehicle combination.

[0017] This invention constructs a dual optimization objective that includes minimizing the capacity gap and minimizing transportation costs, and uses an integer linear programming algorithm to solve it. This achieves refined and optimized vehicle matching, and can find the best balance between vehicle loading efficiency and transportation costs while meeting delivery needs, thereby improving resource utilization.

[0018] Preferably, the step of planning the vehicle route for the optimal vehicle combination within each delivery zone includes: taking the order destination within the delivery zone and the optimal vehicle combination matched for it as input, and solving the problem using a vehicle route planning algorithm.

[0019] Preferably, the step of combining the initial cluster center and the road network distance to cluster the multiple order destinations to obtain multiple delivery partitions includes: using the initial cluster center as the initial centroid of the K-Means clustering algorithm; in each iteration of the K-Means clustering algorithm, based on the road network distance, assigning each order destination to the cluster to which the nearest centroid belongs; until the K-Means clustering algorithm converges to obtain the multiple delivery partitions.

[0020] Preferably, obtaining the road network distance and delivery route from multiple order destinations to the preset starting point includes: calling the application programming interface of the map service, using the geographical coordinates of the preset starting point as the starting point and the geographical coordinates of each order destination as the ending point, requesting and receiving the driving route and its corresponding route length, recording the driving route as the delivery route, and recording the driving route and its corresponding route length as the road network distance.

[0021] In a second aspect, the present invention provides an agricultural resource scheduling system based on smart villages. The agricultural resource scheduling system based on smart villages includes a memory and a processor. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the agricultural resource scheduling method based on smart villages of the present invention is implemented.

[0022] By adopting the above technical solution, a computer program for agricultural resource scheduling based on smart villages, as described in the first aspect of this invention, is generated and stored in a memory for loading and execution by a processor. This allows for the creation of a terminal device based on the memory and processor, facilitating its use.

[0023] The beneficial effects of this invention are as follows: Addressing the challenges of zoning and scheduling in agricultural resource logistics distribution, this invention designs a phased optimization scheme: In the order zoning stage, statistical analysis is conducted based on the distribution characteristics of road network distances. By analyzing the road network distances from all order destinations to warehouses, geographically concentrated order clusters are identified. Further filtering is performed based on the similarity of delivery routes. By evaluating the overlap and directional consistency of order delivery routes within the region, invalid areas with similar distances but significant route divergences are removed. This two-stage filtering process determines high-quality initial cluster centers, ensuring that zoning conforms to road network conditions and that subsequent delivery route planning is reasonable, thus improving zoning feasibility. In the vehicle scheduling stage, an integer linear programming model that balances loading efficiency and transportation costs is constructed. Combining the total order weight, maximum delivery distance, and vehicle parameters of each delivery zone, a suitable vehicle combination is dynamically matched for each delivery zone, reducing additional costs caused by wasted capacity or improper vehicle selection. Ultimately, this achieves the dual optimization goals of improving overall scheduling efficiency and controlling costs. Attached Figure Description

[0024] Figure 1 A flowchart illustrating an agricultural resource scheduling method based on smart villages, provided as an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram illustrating a scenario where agricultural resource scheduling is applied according to an embodiment of the present invention.

[0026] Figure 3 A schematic diagram of the partitioning results of an agricultural resource scheduling method based on smart villages provided in an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the delivery zoning results obtained using a traditional Euclidean distance-based clustering algorithm;

[0028] Figure 5 This is a structural block diagram of an agricultural resource scheduling system based on smart villages, provided as an embodiment of the present invention. Detailed Implementation

[0029] The first aspect of this invention provides a method for agricultural resource scheduling based on smart villages, such as... Figure 1 As shown, the method includes steps S100-S500:

[0030] Step S100: Obtain the road network distance and delivery route from multiple order destinations to the preset starting point.

[0031] It should be noted that the inefficiency of existing order partitioning technologies stems from their unrealistic distance measurement methods, employing straight-line distances such as Euclidean distance for calculations. Therefore, this invention aims to construct a data foundation reflecting real-world road networks, using actual driving routes and distances as the basis for all subsequent calculations, thus providing data support for the rationality of delivery partitioning and route planning. Figure 2 The diagram shown is a schematic representation of a scenario in which this invention is applied. This scenario is a typical example of smart rural agricultural resource distribution, such as fertilizers and seeds. The diagram includes an agricultural resource warehouse as the distribution starting point, multiple geographically dispersed order destinations, and a river as a natural geographical barrier. The horizontal axis represents longitude, and the vertical axis represents latitude, clearly showing the distribution of these geographical entities. The data to be obtained in this step is based on... Figure 2 The geographic entities shown and their topological relationships.

[0032] Specifically, the first step is to access the enterprise resource planning (ERP) of agricultural resource logistics companies. The system retrieves and extracts all pending order information. For each order, it parses out the geographic coordinates of its final destination, formatted as longitude-latitude decimal coordinates to ensure consistency with the map API coordinate system.

[0033] Subsequently, the application programming interface of a commercial map route planning service, such as Baidu Maps API or Gaode Maps API, is called to request and obtain the shortest driving route between the warehouse and each order destination based on the actual road network, with the warehouse's geographical coordinates as the starting point and each order destination's geographical coordinates as the ending point. The driving route is recorded as the delivery route, and the driving route and its corresponding route length are recorded as the road network distance.

[0034] As a preferred implementation, path data can be stored as a series of continuous geographic coordinates, while distance data is stored as floating-point values. These data together form the basis for subsequent analysis steps.

[0035] At this point, the delivery routes and road network distances of all order destinations relative to the warehouse have been obtained.

[0036] Step S200: Based on the distribution characteristics of the road network distance, candidate cluster centers are determined, and the candidate cluster centers are screened in combination with the similarity of the delivery routes to obtain the initial cluster centers.

[0037] It's important to note that traditional clustering algorithms typically select initial cluster centers randomly during partitioning, then iterate and optimize multiple times. This process is not only computationally intensive and time-consuming, but the randomness can also lead to convergence to local optima, affecting the quality of the final partitioning. Especially during peak sales seasons, agricultural resource demand is often clustered at the township level, meaning orders geographically form several clusters at varying distances from warehouses. Utilizing this characteristic, statistical analysis can be used to locate the centers of these clusters, serving as high-quality initial cluster centers, thereby improving the efficiency and accuracy of the partitioning algorithm.

[0038] Specifically, firstly, the road network distances from all order destinations to the warehouse are statistically analyzed, and a distance histogram is constructed. By constructing the distance histogram, one or more peak distance intervals can be identified, and each peak interval represents a potential order cluster area.

[0039] Considering that agricultural resource sales points in townships are usually distributed in relatively concentrated areas, an appropriate distance interval can be selected for statistical analysis. In this embodiment, the distance interval is preferably 5 kilometers. If the interval is too large, such as more than 10 kilometers, orders from different townships may be merged for statistical analysis, reducing the representativeness of the peak interval. If the interval is too small, the histogram may be too sparse and it may be difficult to form obvious peaks. Therefore, using an interval of 5 kilometers can achieve a good balance between the effectiveness of data statistics and geographical differentiation.

[0040] Secondly, for each identified peak distance interval, the arithmetic mean of the geographic coordinates of all order destinations within that interval is calculated to obtain the distance mean point. Then, the geographic coordinates of the actual order destination closest to this distance mean point are selected as the candidate cluster center corresponding to that peak distance interval. This process is repeated to find corresponding candidate cluster centers for all peak distance intervals.

[0041] It should be noted that in actual geographical distribution, there may be two different towns that, although geographically close, are located in different directions from the warehouse. In such cases, candidate cluster centers selected solely based on distance histograms may not be representative enough. Therefore, this invention introduces path similarity evaluation to differentiate between these situations. Orders within the same town typically have a high degree of overlap in their delivery routes; however, orders from different towns, even if geographically close, may have their paths diverge early on. To address this, this invention calculates the selectivity of each candidate cluster center, which is an indicator of the path similarity within the order cluster it represents.

[0042] Specifically, for the first For each peak distance interval and its corresponding candidate cluster center, calculate the similarity between the delivery paths of any two orders within that distance interval, and calculate the average similarity of all order pairs. The more similar the paths are, the higher the cohesion of the cluster and the higher the selectivity of its candidate centers.

[0043] Based on the above logic, selectivity Satisfying the relation:

[0044] ;

[0045] in, It is the first The selectivity of candidate cluster centers at peak distance intervals; It is the first The total number of orders contained within each peak distance interval; , They are the first The, the The delivery path of an order from the warehouse to the order destination can be viewed as a sequence of coordinate points. There are two delivery routes. and The path difference is calculated by considering that the Dynamic Time Warping (DTW) algorithm can measure the similarity between two sequences of different lengths and is very suitable for path comparison. In this embodiment, the Dynamic Time Warping distance (DTW distance) is preferred for calculation. It is a natural exponential function.

[0046] In this formula, the denominator This represents the total number of order pairs, the entire fractional part. The average DTW distance between all order paths within the peak interval was calculated. When orders are geographically highly clustered and paths have a high degree of overlap, the average DTW distance is small, causing the exponential function to approach a power value of 0. Thus making The value approaches A high DTW distance indicates a high degree of selectivity; conversely, if orders are distributed in different directions, resulting in significant path divergence and a large average DTW distance, then... The value approaches This indicates low selectivity.

[0047] It should be added that when If there is only one order within a cluster, its internal path similarity should be considered the highest. Therefore, selectivity can be set in this case. .

[0048] Finally, after calculating the selectivity of all candidate cluster centers... Furthermore, considering that methods such as Otsu's method, which maximize the ratio of between-class variance to within-class variance, can automatically determine the optimal segmentation threshold without prior knowledge, and are particularly suitable for binary classification of continuously distributed index values, effectively avoiding subjective bias caused by manually setting thresholds, this invention employs Otsu's method for all... The values ​​are processed to determine the optimal optionality splitting threshold. All optionalities are then processed. Candidate cluster centers that exceed the selectivity threshold are marked as the final, highly representative initial cluster centers. The Otsu method is existing technology and will not be elaborated upon here.

[0049] Thus, an initial set of cluster centers that can reflect the actual geographical cluster distribution was obtained.

[0050] Step S300: Combine the initial cluster centers and road network distances to cluster the multiple order destinations to obtain multiple delivery zones.

[0051] It should be noted that after obtaining the initial cluster centers, order partitioning can be performed. This step does not use the iterative logic based on Euclidean distance in traditional methods, but uses distance metrics and partitioning rules that fit the real road network to carry out iterative calculations, ensuring that the partitioning results are feasible for actual delivery scenarios.

[0052] Specifically, the geographical coordinates of all order destinations are merged to form a list containing... The algorithm uses a set of points representing fixed destinations as initial centroids for iterative self-organizing clustering algorithms such as K-Means. In each iteration, for each order destination in the set, the road network distance to each cluster center is calculated, and the destination is assigned to the cluster containing the nearest center. After all points are assigned, the centroid of each cluster is recalculated; for example, the point with the smallest sum of road network distances from all points within the cluster is selected as the new centroid. This process is repeated until all cluster centers no longer change, indicating algorithm convergence. After convergence, the final optimal clustering result is obtained. Each cluster represents an independent delivery partition, and all order destinations within the cluster together constitute the set of points for that delivery partition.

[0053] like Figure 4As shown in the figure, the delivery partitioning result is obtained using a traditional Euclidean distance-based clustering algorithm. The figure includes an agricultural resource warehouse as the delivery origin, multiple geographically dispersed order destinations, and a river as a natural geographical barrier. In addition, the figure contains multiple delivery partitions. The horizontal axis represents longitude, and the vertical axis represents latitude. It can be seen that because the traditional algorithm cannot detect the existence of the river, order destinations located on both sides of the river are incorrectly assigned to the same delivery partition. Such a partitioning scheme requires vehicles to take long detours during actual delivery, resulting in low efficiency.

[0054] In contrast, such as Figure 3 The partitioning results obtained using the method described in this invention also include the agricultural resource warehouse as the delivery starting point, multiple geographically dispersed order destinations, and a river as a natural geographical barrier. In addition, the figure contains multiple partitions, with the horizontal axis representing longitude and the vertical axis representing latitude. It can be seen that by employing road network distance-based clustering, this invention successfully divides the order set into multiple geographically connected and compact partitions, such as partition 1, partition 2, and partition 3. None of the partitions cross the river, demonstrating that this invention can generate reasonable and efficient delivery partitions.

[0055] At this point, the reasonable partitioning of all orders has been completed, and the order set for each partition has been obtained.

[0056] Step S400: Based on the total order weight and the furthest road network distance in each delivery zone, and in conjunction with preset vehicle cost parameters, match the optimal vehicle combination between vehicle loading utilization and estimated transportation cost for each delivery zone.

[0057] It should be noted that, in order to solve the problem that the mismatch between transport capacity and cargo volume may be caused by the single vehicle assignment mode, resulting in wasted transport capacity or increased costs, this invention proposes to dynamically match the optimal vehicle combination for each delivery zone, so as to maximize the loading efficiency of vehicles while meeting the transport demand, thereby minimizing transportation costs.

[0058] Specifically, starting from The system retrieves the types of all available vehicles, the available quantity of each type, and their corresponding rated load capacity; then, for the order set of each delivery zone, it aggregates and calculates the total weight of the agricultural products contained therein. (The last sentence appears to be incomplete and possibly refers to a different system or process.) Taking a delivery zone as an example, let its total weight be... The problem of finding the optimal vehicle combination for this delivery zone can be constructed as an integer linear programming model, the goal of which is to ensure that the total carrying capacity is not less than [a certain value]. Under the constraints, the total carrying capacity of the selected vehicle combination is equal to... The difference is the smallest.

[0059] Based on the above logic, for the first The objective function for finding the optimal vehicle combination for each delivery zone satisfies the following relation:

[0060] ;

[0061] in, ;

[0062] The constraint conditions satisfy the following relation:

[0063] ;

[0064] in, It is the first The objective function for the optimal vehicle combination in each delivery zone is given by: It is a decision variable; It is the first The first delivery zone selected The number of each type of vehicle; It is the first The load capacity of each type of vehicle; It is the first Total order weight for each delivery zone; This represents the total number of vehicle types in inventory. It is the first The basic cost of this type of vehicle; It is the first The first delivery zone selected The transportation costs of different types of vehicles; It is the first Cost per kilometer for each type of vehicle; It is the first In the order set of each delivery zone, the maximum value of the shortest actual delivery path distance between the geographical coordinates of the destination of each order and the geographical coordinates of the agricultural resource warehouse, in kilometers; It is the first Constraints on the optimal vehicle combination for each delivery zone; It is the first The inventory quantity of this type of vehicle; It is the set of all non-negative integers; The goal is to find the minimum value.

[0065] In this relation, This reflects the difference between the total load capacity of the selected vehicle combination and the total weight of goods in the delivery zone, calculated by the number of each selected vehicle. Its corresponding load Multiplying and summing the results gives the total carrying capacity. Subtracting the total weight of the goods from this total capacity, the goal of the entire integer linear programming model is to find a set of vehicle numbers. The optimal combination is to minimize the difference between the total load capacity and the total weight, while meeting transportation requirements, thus achieving the highest loading efficiency. This reflects the estimated total transportation cost of the vehicle combination selected for the current delivery zone, consisting of the fixed base cost generated by each type of vehicle and the total mileage cost estimated based on the cost per kilometer of the vehicle and the delivery distance of the farthest order within the delivery zone. The smaller this value, the lower the estimated operating cost required to complete the delivery tasks in this delivery zone.

[0066] The goal of this integer linear programming model is to find a set of integer solutions. , so that the objective function Minimum. The first constraint ensures that the total carrying capacity is sufficient to transport all goods; the second constraint ensures that the number of selected vehicles does not exceed the available number; the third constraint specifies that the decision variable is a non-negative integer.

[0067] It should be noted that this optimization model only applies to the total weight of orders. Delivery zones, for For empty delivery zones, no vehicle matching is required, so there is no need to substitute them into the formula for calculation.

[0068] For example, suppose the first Total weight of each delivery zone for Tonnage, delivery range for orders within the designated area up to 100 kilometers, and three types of vehicles available in the fleet: Type 1: Heavy-duty tons, available Vehicles, basic costs Yuan, cost per kilometer Yuan / km, according to The calculation formula for its transportation cost =250 yuan; Type 2: Load capacity tons, available Vehicle, basic cost Yuan, cost per kilometer Yuan / km, according to The calculation formula for its transportation cost =450 yuan; Type 3: Load capacity tons, available Vehicle, basic cost Yuan, cost per kilometer Yuan / km.

[0069] Its corresponding objective function is The corresponding constraints are: ;and ;and , , It is a non-negative integer.

[0070] Based on the objective function and constraints, an enumeration method can be used to iterate through all combinations of vehicle numbers that satisfy the constraints, calculate the objective function value, and find the combination corresponding to the minimum value. By solving this integer programming problem through enumeration, the optimal solution can be obtained. , , At this point, the total load capacity is tons, the first optimization objective is The second optimization objective is The overall objective function is 1.857. Although this combination has a 1-ton load waste, its objective function value is minimized, which means that the optimal state is achieved in the balance between cost control and waste control. Conversely, if only the carrying efficiency is the goal, such as pursuing zero load waste, the transportation cost will be increased due to the limited choice of vehicle combination.

[0071] After determining the optimal vehicle mix for the delivery zones, the number of available vehicles in the fleet needs to be updated; for example, the inventory of type 2 vehicles should be updated from [previous data]. Vehicle becomes Then, repeat the above optimization process for the next delivery zone until all delivery zones are matched with the optimal vehicle combination.

[0072] At this point, the optimal vehicle combination for each delivery zone has been obtained.

[0073] Step S500: Perform vehicle route planning for the optimal vehicle combination within each distribution zone to obtain the driving routes of each vehicle, thereby completing the scheduling of agricultural resources.

[0074] It should be noted that, through the aforementioned steps, the original complex global optimization problem covering all orders and vehicles has been successfully decomposed into several smaller, independent sub-problems at the delivery zone level, effectively reducing computational complexity and workload. After completing the matching of zones and vehicle combinations, this invention needs to solve the core delivery execution problem for each delivery zone, which is the vehicle routing problem (VRP) that needs to take into account the differences in vehicle models, load limits, maximum vehicle mileage, and order delivery time. Based on this, for each independent delivery zone, the constrained VRP problem is further solved, ultimately planning an efficient delivery route that satisfies all constraints for each vehicle already determined within the zone.

[0075] Specifically, for the first For each delivery zone, an input instance of the VRP problem is first constructed. Key parameters are derived from the output of previous steps and the data obtained in step S100: these include the zone order set, optimal vehicle combination, warehouse geographical location, and road network distances or travel times between customer points and between customer points and warehouses, etc. Then, a vehicle routing algorithm that balances solution efficiency and solution quality is selected to solve the instance. In this embodiment, the preferred metaheuristic algorithm is the Genetic Algorithm (GA). Genetic algorithms are existing technology and will not be elaborated upon here.

[0076] During the algorithm's iterative optimization phase, the path schemes are encoded as individuals, and the optimal path combination that satisfies the constraints and minimizes the total travel distance is searched iteratively using operations such as selection, crossover, and mutation. After the algorithm finishes running, the output optimal individual is the best vehicle route scheme for the corresponding delivery zone, clearly indicating the order in which each vehicle in the zone should visit which customer points.

[0077] It should be added that implementers can adjust the solution objective to minimize the total delivery time or minimize the total cost according to actual needs, or they can use well-known VRP solution algorithms such as simulated annealing and ant colony optimization.

[0078] After processing by the VRP algorithm, an optimal delivery route will be output for each assigned vehicle in the partition. This route specifies the order of order access and ensures that the total driving distance or total time is minimized while satisfying all constraints.

[0079] At this point, the intelligent scheduling of all agricultural resources has been completed, and the optimal driving route has been planned for each vehicle involved in delivery.

[0080] The second aspect of this embodiment provides an agricultural resource scheduling system based on smart villages, such as... Figure 5 As shown, the agricultural resource scheduling system based on smart villages includes a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the first aspect of the present invention, an agricultural resource scheduling method based on smart villages, is implemented.

[0081] The agricultural resource scheduling system based on smart villages also includes other components well known to those skilled in the art, such as communication buses and communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0082] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (DRAM), high-bandwidth memory, hybrid memory cube, etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0083] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A wisdom-village-based agricultural resource scheduling method, characterized in that, Including the following steps: Obtain the road network distance and delivery route from multiple order destinations to the preset origin; Candidate cluster centers are determined based on the distribution characteristics of the road network distance, and the candidate cluster centers are screened based on the similarity of the delivery routes to obtain the initial cluster centers. By combining the initial cluster centers and road network distances, the multiple order destinations are clustered to obtain multiple delivery zones; Based on the total order weight and the furthest road network distance within each delivery zone, and combined with preset vehicle cost parameters, the optimal vehicle combination between vehicle loading utilization and estimated transportation cost is matched for each delivery zone. Vehicle route planning is performed on the optimal vehicle combination within each distribution zone to obtain the driving routes of each vehicle, thereby completing the scheduling of agricultural resources. Determining candidate cluster centers based on the distribution characteristics of the road network distances includes: statistically analyzing the road network distances from all order destinations to the preset starting point and constructing a distance histogram; determining the peak intervals in the distance histogram as clustering regions, and selecting an order destination as a candidate cluster center within each clustering region; The candidate cluster centers are screened based on the similarity of the delivery routes to obtain the initial cluster centers, including: for each candidate cluster center, the set of order destinations in its corresponding cluster area is recorded as a candidate partition, and the average path similarity within each candidate partition is calculated; the candidate cluster centers with an average path similarity greater than a preset similarity threshold are recorded as the initial cluster centers. 2.The smart village-based agricultural resource scheduling method according to claim 1, characterized in that, The step of selecting an order destination as a candidate cluster center within each cluster region includes: Calculate the average of the geographic coordinates of all order destinations within the clustered area to obtain the geographic center point; The order destination closest to the geographic center point within the clustering region is selected as the candidate cluster center. 3.The smart rural-based agricultural resource scheduling method according to claim 1, characterized in that, The calculation of the average path similarity within each candidate partition includes: using a dynamic time warping algorithm to calculate the similarity between any two delivery paths within the candidate partition. 4.The smart rural-based agricultural resource scheduling method according to claim 1, characterized in that, The acquisition of the optimal vehicle combination includes: For each delivery zone, calculate the difference between the total rated load of the vehicle combination and the total weight of the order, and denote it as the first optimization objective; Based on preset vehicle cost and variable cost parameters, combined with the farthest road network distance of the delivery zone, the unit transportation cost of each available vehicle is calculated, and the sum of the unit transportation costs of all vehicles in the vehicle combination is calculated and recorded as the second optimization objective. Construct an objective function that optimizes the index based on the first and second optimization objectives; The objective function is solved using an integer linear programming algorithm to obtain the optimal vehicle combination. 5.The smart village-based agricultural resource scheduling method according to claim 1, characterized in that, The process of planning vehicle routes for the optimal vehicle combination within each delivery zone includes: The optimal vehicle combination matching the order destination within the delivery zone is used as input, and a vehicle route planning algorithm is employed to solve the problem. 6.The smart-village-based agricultural resource scheduling method according to claim 1, characterized in that, The process combines the initial cluster centers and road network distances to cluster the multiple order destinations, resulting in multiple delivery zones, including: The initial cluster centers are used as the initial centroids for the K-Means clustering algorithm; In each iteration of the K-Means clustering algorithm, each order destination is assigned to the cluster of the nearest centroid based on the road network distance; The K-Means clustering algorithm continues until it converges, resulting in the multiple delivery partitions. 7.The smart rural-based agricultural resource scheduling method according to claim 1, characterized in that, The process of obtaining the road network distance and delivery route from multiple order destinations to the preset origin includes: The application programming interface of the map service is invoked, with the geographic coordinates of the preset starting point as the starting point and the geographic coordinates of each order destination as the ending point, to request and receive the driving route and its corresponding route length, and the driving route is recorded as the delivery route, and the driving route and its corresponding route length are recorded as the road network distance.

8. An agricultural resource scheduling system based on smart village, characterized in that, The smart village-based agricultural resource scheduling system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the smart village-based agricultural resource scheduling method according to any one of claims 1-7 is implemented.

Images