Multi-modal collaborative distribution scheduling system and method

Abstract

The invention provides a multi-modal collaborative delivery scheduling method, which comprises the following steps of: constructing a dynamic environment model according to topographic data and traffic condition data, extracting delivery demand characteristics from real-time order information, and splitting a complex order into a plurality of sub-task units by adopting a task decomposition technology to obtain a decomposed sub-task set; extracting sub-task features of adjacent areas and similar time windows from the optimized sub-task sequence, performing clustering analysis on the sub-tasks by adopting an intelligent combination technology, determining space-time relevance among the sub-tasks, and generating a preliminary sub-task aggregation group; and calculating the path cost and the time cost of each delivery batch according to the final resource allocation scheme and the sub-task aggregation grouping, and carrying out optimization iteration on the batch path by adopting a genetic algorithm to obtain a time table and a route plan of the efficient delivery batch.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a multimodal collaborative distribution scheduling system and method. Background Art

[0002] Problem background:

[0003] The field of distribution resource management occupies a vital position in the modern logistics system, which is directly related to the improvement of logistics efficiency, cost control and customer experience.

[0004] With the rapid growth of e-commerce and instant delivery demand, how to efficiently integrate multiple delivery resources has become a key issue in the development of the industry.

[0005] The rise of unmanned delivery technologies such as drones and driverless cars has injected new vitality into the logistics industry, but it has also brought complexity to resource coordination and task optimization.

[0006] Current solutions mostly focus on the management of a single delivery method, such as manual delivery or scheduling of unmanned vehicles, and lack unified control and collaborative optimization of multimodal delivery resources.

[0007] This fragmented management approach is unable to cope with diverse delivery scenarios, resulting in low resource utilization and limited delivery efficiency.

[0008] The limitation of existing methods is that they are often unable to flexibly adapt to different terrain characteristics, traffic conditions and the heterogeneity of distribution resources.

[0009] Traditional scheduling systems mostly rely on static rules or simple algorithms, which makes it difficult to achieve dynamic decomposition and efficient aggregation of tasks. Especially when faced with complex delivery needs, sub-tasks are allocated in a scattered manner and lack intelligent integration, resulting in a waste of time and resources.

[0010] The core challenges focus on three technical factors: first, the centralized management and coordination mechanism of multimodal distribution resources is not yet mature; second, the task decomposition technology for complex scenarios lacks adaptability; and third, the optimization methods for subtask aggregation are not sufficient to cope with the combined demands of similar areas or similar times.

[0011] Because these technical factors have not been effectively resolved, the contradiction of idle resources and accumulated tasks often occurs during the distribution process. Especially during peak hours or in areas with complex terrain, it is difficult to generate efficient delivery batches, which in turn affects the overall logistics efficiency.

[0012] Therefore, how to develop a unified operation control platform for multi-modal distribution resources, achieve centralized management of drones, unmanned vehicles, and manual distribution, and through the two-way optimization technology of task decomposition and aggregation, intelligently combine subtasks on the basis of dynamically adapting to terrain and traffic conditions to form efficient distribution batches has become the key problem that needs to be solved urgently in this research. Summary of the Invention

[0013] The present invention provides a multi-modal collaborative distribution scheduling method, which mainly includes:

[0014] Obtain the terrain data, traffic condition data, and real-time order information within the distribution area, identify the available resource range of the unmanned distribution technology and manual distribution collaboration through the multi-modal distribution mode, and determine the initial state and location distribution of each resource in the current period;

[0015] Construct a dynamic environment model according to the terrain data and traffic condition data, extract the distribution demand characteristics from the real-time order information, and use the task decomposition technology to split the complex order into several subtask units to obtain the decomposed subtask set;

[0016] For the decomposed subtask set, obtain the geographical location, time window, and resource demand attributes of each subtask, match the terrain and traffic conditions through the dynamic adaptability algorithm, and judge that if the subtask conflicts with the current environmental conditions, then adjust the priority and execution order of the subtask to obtain the optimized subtask sequence;

[0017] Extract the subtask characteristics of adjacent regions and similar time windows from the optimized subtask sequence, perform clustering analysis on the subtasks using the intelligent combination technology, determine the spatio-temporal correlation between the subtasks, and generate a preliminary subtask aggregation grouping;

[0018] Obtain the total resource demand and distribution path constraints of the preliminary subtask aggregation grouping, allocate the resource combination of the unmanned distribution technology and manual distribution collaboration through the resource centralized management module, and judge that if the resource demand exceeds the current capacity, then allocate and supplement from the standby resource pool to obtain the final resource allocation plan;

[0019] Calculate the path cost and time cost of each distribution batch according to the final resource allocation plan and the subtask aggregation grouping, and use the genetic algorithm to optimize and iterate the batch path to obtain the schedule and route plan of the efficient distribution batch;

[0020] Update the real-time state of the unmanned distribution technology and manual distribution collaboration through the schedule and route plan of the efficient distribution batch, collect feedback data from the distribution execution process, and judge that if the feedback data indicates resource idleness or task backlog, then trigger the dynamic adaptability algorithm to readjust the task decomposition and aggregation to obtain a new batch scheduling result;

[0021] After obtaining the new batch scheduling result, the resource centralized management module synchronously updates the operating status and location information of all distribution resources, generates the task decomposition and aggregation input data for the next cycle, and determines the continuous optimization ability of the system in the dynamic environment.

[0022] The present invention also provides a multi-modal collaborative distribution scheduling system, which is characterized by including:

[0023] Multi-modal data acquisition module: configured to acquire terrain data, traffic condition data, and real-time order information within the distribution area, process through a geographic information system to obtain the distribution of accessible paths, combine the K-means algorithm for clustering analysis of order-dense areas and resource demand distribution, identify the available resources range for unmanned distribution and the collaborative resource data for manual distribution, determine the initial states, location distributions, and collaborative resource allocation schemes of unmanned distribution devices and manual distribution, and calculate the optimal path distribution using the Dijkstra algorithm;

[0024] Dynamic environment modeling module: configured to fuse terrain data and traffic conditions to construct a dynamic model, adjust the model through real-time update technology to determine the changing trend of traffic conditions, extract distribution demand characteristics from order information and perform priority ranking, use task decomposition technology to split complex orders into a set of subtasks, and optimize the subtask division and verify the task execution sequence through data fusion technology;

[0025] Task decomposition and optimization module: configured to obtain the geographical location, time window, and resource demand attributes of subtasks, match terrain and traffic data through a dynamic adaptability algorithm to judge the conflict between tasks and environmental conditions, if there is a conflict, adjust the subtask priorities and execution orders, combine the greedy algorithm to optimize resource allocation, verify the matching consistency of geographical location and terrain traffic, and generate the finally optimized subtask sequence;

[0026] Subtask clustering and aggregation module: configured to extract the adjacent area and time window characteristics of subtasks, perform clustering analysis using intelligent combination technology and the K-means algorithm, determine the spatio-temporal correlation between subtasks, and generate refined subtask aggregation groups;

[0027] Resource centralized management module: configured to obtain the total resource demand and distribution path constraints of subtask aggregation groups, integrate the resource combinations of unmanned distribution technology and manual distribution collaboration, if the resource demand exceeds the current capacity, allocate and supplement from the standby resource pool, optimize through the linear programming algorithm to obtain the final resource allocation scheme, and monitor the execution status of the scheme in real time;

[0028] Path and time optimization module: configured to calculate the path cost and time cost according to the resource allocation scheme and subtask aggregation groups, use the genetic algorithm to optimize and iterate the batch path and schedule, and generate the schedule and route plan for efficient distribution batches;

[0029] Dynamic feedback adjustment module: Configured to collect distribution execution feedback data through sensors and logging systems. If resource idleness or task backlog is detected, it activates the dynamic adaptability algorithm to readjust task decomposition and aggregation, generates a new batch scheduling result, synchronously updates the operating status and location information of distribution resources, generates the input data for task decomposition and aggregation in the next cycle, and predicts the continuous optimization ability of the system in a dynamic environment through a regression algorithm.

[0030] The technical solution provided by the embodiments of the present invention may include the following beneficial effects:

[0031] The present invention discloses an intelligent scheduling method for collaborative unmanned and manual distribution. It identifies the available resource range through a multi-modal distribution mode, constructs a dynamic environment model, and uses task decomposition technology to split complex orders into subtask units. For the set of subtasks, the present invention applies a dynamic adaptability algorithm to match the terrain and traffic conditions, and adjusts the task priorities and execution sequences. Subsequently, it uses intelligent combination technology to perform clustering analysis on the subtasks to generate a preliminary aggregation grouping. The present invention also allocates unmanned and manual distribution resources through a resource centralized management module, and uses a genetic algorithm to optimize the batch path to obtain an efficient distribution schedule and route plan. During the execution process, the present invention continuously collects feedback data, triggers a dynamic adjustment mechanism to re-optimize task decomposition and aggregation, and realizes the continuous optimization ability in a dynamic environment. This method effectively improves the distribution efficiency and resource utilization rate in complex order scenarios, and provides an intelligent solution for the collaborative mode of unmanned and manual distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flowchart of a multi-modal collaborative distribution scheduling method of the present invention.

[0033] Figure 2 It is a schematic diagram of a multi-modal collaborative distribution scheduling method of the present invention.

[0034] Figure 3 It is another schematic diagram of a multi-modal collaborative distribution scheduling method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] Embodiment 1

[0037] A multi-modal collaborative distribution scheduling system includes:

[0038] Multi-modal data acquisition module: Configured to acquire terrain data, traffic condition data, and real-time order information within the delivery area, process through a geographic information system to obtain the distribution of accessible paths, combine the K-means algorithm for clustering analysis of order-intensive areas and resource demand distribution, identify the scope of available resources for unmanned delivery and collaborative resource data for manual delivery, determine the initial states, location distributions, and collaborative resource allocation plans for unmanned delivery equipment and manual delivery, and use the Dijkstra algorithm to calculate the optimal path distribution;

[0039] Dynamic environment modeling module: Configured to fuse terrain data and traffic conditions to build a dynamic model, adjust the model through real-time update technology to determine the trend of traffic condition changes, extract delivery demand characteristics from order information and prioritize them, use task decomposition technology to split complex orders into a set of subtasks, and optimize subtask partitioning and verify task execution sequences through data fusion technology;

[0040] Task decomposition and optimization module: Configured to obtain the geographical location, time window, and resource demand attributes of subtasks, match terrain and traffic data through a dynamic adaptability algorithm to judge conflicts between tasks and environmental conditions, adjust subtask priorities and execution orders if there are conflicts, optimize resource allocation in combination with the greedy algorithm, verify the matching consistency between geographical locations and terrain traffic, and generate a final optimized subtask sequence;

[0041] Subtask clustering and aggregation module: Configured to extract adjacent area and time window characteristics of subtasks, perform clustering analysis using intelligent combination technology and the K-means algorithm, determine the spatio-temporal correlation between subtasks, and generate refined subtask aggregation groups;

[0042] Resource centralized management module: Configured to obtain the total resource demand and delivery path constraints of subtask aggregation groups, integrate the resource combination of unmanned delivery technology and manual delivery collaboration, allocate and supplement from the backup resource pool if the resource demand exceeds the current capacity, optimize through the linear programming algorithm to obtain the final resource allocation plan, and monitor the execution status of the plan in real time;

[0043] Path and time optimization module: Configured to calculate path costs and time costs based on the resource allocation plan and subtask aggregation groups, use the genetic algorithm to optimize and iterate the batch paths and schedules, and generate schedules and route plans for efficient delivery batches;

[0044] Dynamic feedback adjustment module: Configured to collect distribution execution feedback data through sensors and logging systems. If resource idleness or task backlog is detected, it activates the dynamic adaptability algorithm to re-adjust task decomposition and aggregation, generates a new batch scheduling result, synchronously updates the operating status and location information of distribution resources, generates the input data for task decomposition and aggregation in the next cycle, and predicts the continuous optimization ability of the system in a dynamic environment through a regression algorithm.

[0045] Embodiment 2

[0046] As Figures 1-3 , a multi-modal collaborative distribution scheduling method in this embodiment may specifically include:

[0047] Step S101, obtain the terrain data, traffic condition data, and real-time order information within the distribution area, identify the available resource range of the collaborative operation of unmanned distribution technology and manual distribution through multi-modal distribution mode recognition, and determine the initial state and location distribution of each resource in the current period.

[0048] [1] Obtain the terrain data and traffic condition data within the distribution area, process through a preset geographic information system to obtain the distribution of accessible paths within the area. Combine the real-time order information with the distribution of accessible paths, and use the K-means algorithm for cluster analysis to obtain the order-intensive areas and resource demand distribution. If the traffic condition data in the order-intensive areas exceeds the preset threshold, identify the available resource range of unmanned distribution through multi-modal analysis. Based on the available resource range of unmanned distribution and the distribution of real-time orders, determine the initial state and location distribution of unmanned distribution equipment. Obtain the collaborative resource data of manual distribution, and combine it with the demand distribution in the order-intensive areas to judge the initial state and location distribution of manual distribution. Through the superposition analysis of the location distribution and the initial state, obtain the collaborative resource allocation plan of unmanned distribution and manual distribution in the current period. For the collaborative resource allocation plan, use the Dijkstra algorithm to calculate the optimal path distribution of each resource from the initial state to the target order.

[0049] In a possible implementation, when obtaining the terrain data and traffic condition data within the distribution area, the satellite images and real-time traffic flow information can be integrated through a geographic information system.

[0050] For example, within the distribution area of a certain city, use high-resolution satellite maps to obtain information such as road width and slope, and at the same time analyze the traffic flow density in combination with traffic camera data.

[0051] Preferably, the geographic information system can overlay these data to generate a regional accessible path distribution map, clearly showing the accessible capabilities of main roads, secondary roads, and alleys. This helps to identify which paths are suitable for unmanned distribution equipment to pass through, significantly improving the path planning efficiency.

[0052] Specifically, by combining real-time order information with the distribution of travel paths and using the K-means algorithm for clustering analysis, it is possible to effectively divide order-dense areas.

[0053] For example, there are 100 orders in a certain area at 10 am. The system clusters them into 3 dense areas based on order coordinates and path distribution, with 60% of the orders in the business district. This analysis can intuitively reflect the distribution of resource demands and provide a basis for the allocation of distribution resources.

[0054] It should be noted that the K-means algorithm ensures that order points are assigned to the nearest cluster center through iterative optimization, thereby improving the accuracy of area division.

[0055] In one embodiment, if the traffic condition data in the order-dense area exceeds a preset threshold, such as the speed of the main road in the business district being lower than 10 km / h, then multi-modal analysis is initiated to identify available resources for unmanned delivery.

[0056] For example, the system integrates data on drones, unmanned vehicles, and smart lockers and confirms that there are 5 drones and 10 lockers available in the business district. This analysis can quickly match resources with demands and avoid traffic congestion from affecting delivery efficiency.

[0057] It can be understood that the accurate identification of the scope of unmanned delivery resources helps to reduce the delivery delay rate.

[0058] For example, when determining the initial state and location distribution of unmanned delivery equipment based on the scope of unmanned delivery resources and order distribution, drones can be preferentially deployed near lockers close to the order-dense area. Suppose there are 3 lockers in the business district, with 2 drones deployed at each location and the initial state set to standby. This distribution can shorten the response time and improve the equipment utilization rate.

[0059] Preferably, the system will also dynamically adjust the equipment state according to the order priority to ensure that high-priority orders are processed first.

[0060] In a possible implementation, when obtaining data on collaborative resources for manual delivery, the real-time location and capacity of delivery personnel can be analyzed in combination with the demands in the order-dense area.

[0061] For example, the system detects that an additional 10 delivery personnel are needed in the business district, 5 of whom are currently on standby nearby, and the other 5 can be deployed from neighboring areas.

[0062] It should be noted that this judgment of the initial state and location distribution can balance the collaborative efficiency of manual and unmanned delivery and avoid resource waste.

[0063] Specifically, when generating a collaborative resource allocation plan through the overlay analysis of the location distributions of unmanned delivery and manual delivery, drones can be preferentially assigned to traffic-congested areas, while manual delivery personnel are responsible for areas with complex terrains.

[0064] For example, the main roads in the commercial area are congested. Drones are responsible for orders from high-rise office buildings, and delivery personnel are responsible for orders from old residential areas. This plan can give full play to the advantages of both and improve the overall delivery efficiency.

[0065] In one embodiment, for the collaborative resource allocation plan, the Dijkstra algorithm is used to calculate the optimal path distribution.

[0066] For example, when a drone departs from a locker to an order point, the system calculates the shortest flight path based on the path length and obstacle avoidance requirements; the delivery personnel plan the walking or electric vehicle path according to the road passability and the distance to the order point.

[0067] It can be understood that this path optimization can significantly reduce the delivery time and energy consumption.

[0068] For example, an order point is 500 meters away from the locker. It takes 2 minutes for the drone to fly directly, while it takes 10 minutes for the delivery personnel to take a detour. After optimizing the path, the overall efficiency is increased by 30%. Through precise calculation, this method ensures the efficient use of resources and brings higher stability and customer satisfaction to the delivery business.

[0069] Step S102: Construct a dynamic environment model based on terrain data and traffic condition data, extract the delivery demand characteristics from the real-time order information, and use task decomposition technology to split complex orders into several sub-task units to obtain the decomposed sub-task set.

[0070] Construct a dynamic model through the fusion of terrain data and traffic conditions to obtain an environmental feature description. Use real-time update technology to adjust the dynamic model to determine the change trend of traffic conditions. Extract the delivery demand characteristics from the order information to obtain the demand priority ranking. Perform task decomposition according to the demand priority ranking to obtain a preliminary sub-task division. If there are overlaps in the sub-task division, adjust the sub-task set through data fusion technology to judge the optimized sub-task set. Divide the unit according to the optimized sub-task set to obtain the final delivery task allocation. Verify the final delivery task allocation through a machine learning algorithm to determine the task execution sequence.

[0071] Specifically, construct a dynamic model through the fusion of terrain data and traffic conditions to obtain an environmental feature description.

[0072] For example, within the delivery area, the terrain data includes information such as slope and road surface type, and the traffic conditions include traffic flow and intersection congestion level.

[0073] It is understandable that when integrating these data, the terrain data can be first converted into a two-dimensional grid, with each grid marked with height and road surface attributes, and then the real-time heat map of traffic conditions is overlaid to form a dynamic model containing environmental features.

[0074] Exemplarily, on the main road of a certain area, it is flat but has a large traffic flow, and the model will mark it as an area with high traffic costs. While the small road is narrow but unobstructed, it is marked as a low-cost path. This model intuitively reflects the impact of the environment on distribution and facilitates subsequent decision-making. The dynamic model is adjusted using real-time update technology to determine the changing trend of traffic conditions.

[0075] Specifically, traffic information can be updated every 5 minutes through roadside sensors or navigation data.

[0076] In one embodiment, due to the increase in early morning peak traffic flow on a certain road section, the model will dynamically increase the traffic cost of this road section and predict the congestion trend in the next 30 minutes.

[0077] Preferably, combined with historical data analysis, the model can also identify regular changes, such as fixed congestion points from 7 to 9 o'clock on weekdays. This real-time adjustment ensures that distribution decisions are always based on the latest environmental information. Extract the distribution demand characteristics from the order information and obtain the priority ranking of the demands.

[0078] For example, the order data includes the delivery time window, the weight of the goods, and the customer type. One possible implementation is to set the priority of urgent orders (such as fresh food) as the highest, and that of ordinary orders as the second.

[0079] Exemplarily, a customer places an order requiring delivery within 1 hour, and the system will automatically raise its priority and mark it as a high-time-efficiency demand. This ranking method clearly divides the urgency of tasks and provides a basis for resource allocation. Perform task decomposition according to the demand priority ranking to obtain a preliminary sub-task division.

[0080] It is understandable that high-priority orders can be decomposed into independent sub-tasks, and low-priority orders can be combined for processing.

[0081] In one embodiment, among 10 orders, 3 are of high priority, and the system assigns them to separate deliverymen, and the remaining 7 are combined into one regional task. This decomposition reduces resource waste and improves the response speed to critical orders. If there are overlaps in the sub-task division, adjust the sub-task set through data fusion technology to judge the optimized sub-task set.

[0082] Specifically, overlap means that multiple sub-tasks cover the same area or time window.

[0083] Exemplarily, when two deliverymen are both assigned tasks in the same community, the system will merge them into one task by integrating the order addresses and time requirements, and reallocate resources. This adjustment avoids duplicate deliveries and improves overall efficiency. According to the optimized subtask set division unit, the final delivery task allocation is obtained.

[0084] For example, the system divides the optimized tasks into the smallest execution units according to regions and time windows.

[0085] In one embodiment, there are 5 subtasks in a certain region, and the system assigns them to 3 deliverymen, with each person responsible for adjacent units. This division ensures a balanced distribution of tasks and reduces the ineffective movement of deliverymen. The final delivery task allocation is verified through a machine learning algorithm to determine the task execution sequence.

[0086] Preferably, a decision tree model can be used to analyze the rationality of task allocation.

[0087] In one possible implementation, the model inputs the task allocation plan and combines historical delivery data to predict the completion time of each task.

[0088] Exemplarily, a certain allocation plan shows that a deliveryman needs to continuously cross 3 regions. The model prompts that its sequence is inefficient and suggests adjusting it to the order of nearest first. This verification ensures that the task sequence better meets the actual execution requirements.

[0089] Step S103: For the decomposed subtask set, obtain the geographical location, time window, and resource requirement attributes of each subtask, match the terrain and traffic conditions through a dynamic adaptation algorithm, and if a subtask conflicts with the current environmental conditions, adjust the priority and execution order of the subtask to obtain an optimized subtask sequence.

[0090] Obtain the geographical location, time window, and resource requirement attributes in the subtask set, and obtain the terrain, traffic, and environmental condition data through database query. Match the geographical location with the terrain and traffic data through a dynamic adaptation algorithm to determine whether a subtask conflicts with the environmental conditions, and obtain a conflict judgment result. If the conflict judgment result is true, adjust the priority according to the time window to obtain an adjusted priority sequence. According to the adjusted priority sequence, rearrange the execution order to obtain a preliminarily sorted subtask sequence. Analyze the preliminarily sorted subtask sequence through the resource requirement attributes, and use a greedy algorithm to optimize the resource allocation to obtain a resource-optimized subtask sequence. For the resource-optimized subtask sequence, verify the matching consistency between the geographical location and the terrain and traffic to obtain a final optimized sequence. Through the final optimized sequence, update the execution order of the subtask set to determine the task scheduling plan.

[0091] Exemplarily, obtaining the geographical location, time window, and resource requirement attributes in the subtask set can be achieved through database query.

[0092] In a possible implementation, assuming an urban distribution scenario, the subtask set contains multiple distribution points, and the geographical locations are represented as longitude and latitude coordinates. For example, distribution point A is located at 120.5° east longitude and 30.2° north latitude, the time window is from 9:00 to 10:00 in the morning, and the resource requirements are 2 deliverymen and 1 electric vehicle. The database stores the terrain data of the city, such as slope and road width, as well as the real-time traffic conditions, such as the congestion index of road sections. Through SQL queries, these attribute and environmental data can be quickly extracted to ensure a clear data structure for subsequent processing.

[0093] It should be noted that by using a dynamic adaptability algorithm to match the geographical location with the terrain and traffic data to determine whether the subtask conflicts with the environmental conditions, the core lies in environmental adaptability.

[0094] In one embodiment, distribution point A is located in a steep slope area, and the terrain data shows that the slope is 15%, while the maximum climbing ability of the delivery vehicle is 10%. The algorithm detects this conflict and marks it as infeasible. In contrast, distribution point B is located in a flat area with no conflict. This matching ensures the executability of the task by comparing the geographical location of the subtask with the terrain limitations.

[0095] Specifically, if the conflict judgment result is true, the priority is adjusted according to the time window.

[0096] For example, distribution point A cannot be completed on time due to terrain conflict, but its time window is relatively loose and can be adjusted to a lower priority. After adjustment, the priority sequence advances distribution point B in the flat area, and the task with the time window from 9:30 to 10:30 is executed first. This adjustment optimizes the task sorting by analyzing the flexibility of the time window.

[0097] In one embodiment, after rearranging the execution order, a preliminarily sorted subtask sequence is obtained. Assume the sequence is B, C, A, where C is another distribution point in a flat area. Based on the analysis of the resource requirement attributes, both distribution points B and C require 1 vehicle each, while A requires 2 vehicles. The greedy algorithm optimizes the resource allocation and preferentially allocates vehicles to B and C to ensure the maximization of resource utilization.

[0098] Preferably, if the total number of vehicles is 2, the algorithm postpones the execution time of A to generate a sequence with optimized resources.

[0099] It can be understood that the matching consistency between the geographical location and the terrain and traffic is verified to ensure the feasibility of the optimized sequence.

[0100] For example, recheck the sequence B, C, A, confirm that the roads of B and C are unobstructed, and the slope problem of A is solved by replacing the vehicle with higher power. After the final optimized sequence passes, the execution order of the subtask set is updated to form a task scheduling plan.

[0101] For example, delivery point B is completed before 9:30, C is completed at 10:00, and A is scheduled to be executed using a special vehicle at 10:30. This solution is verified through multiple dimensions to ensure efficient scheduling and adaptability to environmental changes.

[0102] Step S104: Extract the sub-task features of adjacent regions and similar time windows from the optimized sub-task sequence, use intelligent combination technology to perform clustering analysis on the sub-tasks, determine the spatio-temporal correlation between the sub-tasks, and generate a preliminary sub-task aggregation grouping.

[0103] Obtain task feature data from the sub-task sequence, process it using feature extraction technology to obtain a feature set. Extract adjacent region and time window information from the feature set to determine the distribution of sub-tasks with similar regions and times. For data with similar regions and times, use intelligent combination technology for grouping to obtain a preliminary clustering result. Perform spatio-temporal correlation analysis on the preliminary clustering result to judge the correlation strength between sub-tasks and determine the optimized clustering grouping. Obtain the optimized clustering grouping, use the K-means algorithm to perform secondary clustering on the sub-task sequence to obtain a refined aggregation grouping. Analyze the matching degree between task features and spatio-temporal correlation through the refined aggregation grouping, judge the grouping consistency, and obtain the final sub-task aggregation grouping. Extract task optimization information from the final sub-task aggregation grouping to determine the execution order and resource allocation plan between sub-tasks.

[0104] Exemplarily, in the logistics distribution task scheduling, obtaining the task feature data of the sub-task sequence is the basis for optimizing the distribution efficiency. The task feature data usually includes information such as the location of the delivery point, the order volume, and the delivery time requirements. Process these data using feature extraction technology, for example, extract spatial distribution features from the delivery point coordinates through clustering analysis, or extract time density features from the order timestamps.

[0105] In a possible implementation, assume that there are 100 delivery points in a city, each with longitude and latitude and order submission time. The feature extraction technology can divide these points into 5 main clusters by region and identify that the peak hours are concentrated from 9 am to 11 am. The resulting feature set provides a structured input for subsequent analysis.

[0106] It should be noted that when extracting adjacent region and time window information from the feature set, the purpose is to identify sub-tasks with similar characteristics.

[0107] For example, based on the spatial distance calculation, delivery points with a distance less than 2 kilometers are regarded as adjacent regions; based on time analysis, orders with overlapping delivery time windows exceeding 30 minutes are regarded as time-similar.

[0108] In one embodiment, 8 out of 10 delivery points in a certain area are located in the same commercial district, and the order time windows are all from 2 pm to 4 pm, indicating that they are highly similar in space and time. This distribution information provides a basis for grouping.

[0109] Specifically, for data that is similar in area and time, intelligent combination technology can group them through a rule engine or a machine learning model.

[0110] Preferably, the rule engine can set priorities, such as preferentially combining subtasks with a short spatial distance and then considering the degree of overlap of the time windows.

[0111] For example, the subtasks of the above 8 delivery points are combined into a group, while the other two remote points are assigned to other groups. This preliminary clustering result reduces the ineffective movement in the delivery.

[0112] In one embodiment, spatio-temporal correlation analysis is used to judge the correlation strength between subtasks. The correlation strength can be measured by the traffic accessibility of the delivery points or the time dependence of the orders.

[0113] For example, the 8 delivery points in the commercial district are connected by a main road, and the traffic time is less than 10 minutes, so the correlation strength is high; while the traffic time of the remote points may exceed 30 minutes, so the correlation strength is low. Based on this, the optimized clustering grouping will preferentially retain the subtask combinations with high correlation degrees to ensure a more compact delivery route.

[0114] It can be understood that using the K-means algorithm to perform secondary clustering on the subtask sequence is to further refine the grouping. The K-means algorithm will reassign the subtasks to more optimal clusters according to task characteristics such as distance, time window, order volume, etc.

[0115] For example, the subtasks in the preliminary grouped commercial district may be subdivided into two clusters, one close to the shopping mall and the other close to the office building. This refined aggregation grouping improves the accuracy of task allocation.

[0116] For example, by analyzing the matching degree of task characteristics and spatio-temporal correlation in the refined aggregation grouping, it can be judged whether the grouping is reasonable.

[0117] In one possible implementation, check whether the average distance of the subtasks within each cluster is less than 1 km and whether the time window overlap rate exceeds 80%. If the distribution of the subtasks within a certain cluster is too scattered, readjust the grouping. This consistency verification ensures the practicality of the grouping.

[0118] Preferably, extract task optimization information from the final subtask aggregation grouping, such as the center point of each cluster as the starting point of the delivery, or determine the preferential delivery order based on the time window.

[0119] In one embodiment, the subtasks in the business district are arranged on the same delivery vehicle, with the starting point set at the shopping center and the sequence arranged from early to late according to the time window. This optimization information is directly converted into a specific execution sequence and resource allocation plan, such as allocating a vehicle with a larger transport volume to high-density areas. This method significantly improves the delivery efficiency.

[0120] Step S105: Obtain the total resource requirements and delivery path constraints of the preliminary subtask aggregation grouping, allocate the resource combination of unmanned delivery technology and manual delivery collaboration through the resource centralized management module, and if the resource requirements exceed the current capacity, allocate supplementary resources from the standby resource pool to obtain the final resource allocation plan.

[0121] Obtain the resource requirements and delivery path constraints of the subtask aggregation grouping through the task decomposition module, determine the preliminary demand distribution data, process the delivery path constraints using the path optimization algorithm to obtain the optimized path allocation plan, integrate the resource combination of unmanned delivery technology and manual delivery collaboration through the resource centralized management module, judge the feasibility of the allocation, if the demand exceeds the current capacity limit, analyze the excess part through the capacity assessment tool, obtain the available resource data of the standby resource pool, allocate supplementary resources from the standby resource pool to the resource combination allocation to obtain the adjusted resource allocation result, optimize the adjusted resource allocation result using the linear programming algorithm to determine the final allocation plan, and verify the execution status of the final allocation plan through the real-time monitoring module to obtain the dynamic adjustment data during the execution process.

[0122] Exemplarily, when obtaining the resource requirements and delivery path constraints of the subtask aggregation grouping through the task decomposition module, the specific task requirements in the unmanned delivery scenario can be analyzed first. Suppose there are 10 delivery points in a certain urban area, and each point requires a different number of packages. For example, point A requires 50 pieces and point B requires 30 pieces. The task decomposition module will organize these points into subtask groups according to the geographical location and the quantity of packages, and at the same time record the delivery path constraints of each group, such as some areas need to be bypassed due to traffic control. Such a decomposition method can clearly present the demand distribution and facilitate subsequent optimization.

[0123] In a possible implementation, when processing the delivery path constraints using the path optimization algorithm, an optimized path allocation plan can be generated based on the actual geographical data.

[0124] For example, for the above 10 delivery points, the algorithm will preferentially select a path with a shorter distance and avoiding the controlled area. Suppose the straight-line distance from point A to point B is 5 kilometers, but due to control, it needs to bypass 7 kilometers. The algorithm will calculate the time and energy consumption costs and recommend an alternative path of 6 kilometers. This method ensures that the path is efficient and meets the actual constraints.

[0125] Specifically, when the resource centralized management module integrates the resource combination of unmanned delivery technology and manual delivery collaboration, it can allocate unmanned vehicles and manual delivery staff according to the task volume.

[0126] For example, among the 50 packages at point A, 40 are delivered by unmanned vehicles, and 10 need to be handled manually due to special requirements. The module will evaluate the load capacity of the unmanned vehicle and the delivery efficiency of the manual staff to determine whether the combination is feasible. If the unmanned vehicle is fully loaded, the module will dynamically adjust the proportion of manual delivery. This integration method can flexibly respond to different task requirements.

[0127] It should be noted that if the demand exceeds the current capacity limit, the capacity assessment tool will analyze the excess part and obtain the data of the backup resource pool.

[0128] For example, if the 50 packages at point A exceed the maximum capacity of 40 packages for the unmanned vehicle, the tool will query the backup resource pool and find an available additional unmanned vehicle or temporary manual delivery staff. The analysis process will consider the scheduling time and cost of the backup resources to ensure that the supplementary resources are in place in a timely manner.

[0129] Preferably, after allocating supplementary resources from the backup resource pool to the resource combination, an adjusted resource configuration result can be obtained.

[0130] For example, after additionally allocating an unmanned vehicle, the delivery demand at point A is met, and the configuration result will be updated to two unmanned vehicles for 40 packages and 10 packages, with the manual delivery staff responsible for the remaining part. This allocation method can quickly respond to sudden demands.

[0131] In one embodiment, when using the linear programming algorithm to optimize the adjusted resource configuration result, the final plan can be determined by simulating different delivery combinations.

[0132] For example, the algorithm will compare the efficiency of fully automatic delivery by unmanned vehicles with the mixed delivery of unmanned vehicles and manual staff, and preferentially select the plan with short time consumption and low cost. Assuming that the mixed delivery can save 20% of the delivery time, the algorithm will lock in this plan. This optimization method can improve the overall delivery efficiency.

[0133] It can be understood that when verifying the execution status of the final allocation plan through the real-time monitoring module, dynamic adjustment data can be obtained.

[0134] For example, during the delivery process, if the unmanned vehicle is delayed at point B due to a temporary road closure, the monitoring module will give real-time feedback and suggest switching to an alternative route or dispatching additional manual delivery staff. The adjustment data will record the delay time and the effect of the solution, providing a reference for subsequent task optimization. This real-time monitoring method can ensure the stability and controllability of the delivery process.

[0135] Step S106: Calculate the path cost and time cost of each delivery batch according to the final resource allocation plan and sub-task aggregation grouping, and use the genetic algorithm to optimize and iterate the batch path to obtain the schedule and route plan of the efficient delivery batch.

[0136] Divide the sub-task groups through the resource allocation plan, calculate the grouping result after task aggregation, and obtain the basic data of the delivery batch. Extract the path cost and time cost from the basic data of the delivery batch, and use the cost calculation method to determine the initial cost value of each batch. For the initial cost value, judge whether the path cost exceeds the preset threshold. If it exceeds, adjust the batch path through the genetic algorithm to obtain the optimized path plan. According to the optimized path plan, obtain the change trend of the time cost, and use the genetic algorithm to iterate the schedule to determine the time series of efficient delivery. Calculate the route plan of each delivery batch through the time series and path plan to obtain the final route distribution data. Obtain the route distribution data and judge whether the time cost meets the preset threshold. If it does not meet, adjust the grouping method of task aggregation and repeat the optimization and iteration process. According to the adjusted grouping method, use the cost calculation method to re-determine the path cost and time cost to obtain the schedule and route plan of the efficient delivery batch.

[0137] Specifically, divide the sub-task groups through the resource allocation plan, calculate the grouping result after task aggregation, and obtain the basic data of the delivery batch. The core is to break down the complex delivery requirements into operable units.

[0138] For example, a city distribution center needs to process 1000 orders per day, including fresh food, daily necessities, etc.

[0139] Preferably, the tasks can be grouped into 10 sub-task groups according to the geographical location and order type, with about 100 orders in each group to ensure balanced grouping. In implementation, the batch data can be automatically generated through the order management system based on the regional division and category priority, such as the fresh food priority batch in the northern region, including 50 orders. The basic data includes the order volume, delivery distance, etc. Extract the path cost and time cost from the basic data of the delivery batch, and use the cost calculation method to determine the initial cost value of each batch. The key is to quantify the delivery cost.

[0140] Specifically, the path cost can be calculated based on the distance and traffic conditions, and the time cost takes into account the delivery duration.

[0141] For example, the total distance of the northern region batch is 200 kilometers, the cost per kilometer is 0.5 yuan on average, and the path cost is 100 yuan. The time cost is calculated based on 15 minutes for each order delivery and an hourly labor cost of 60 yuan, resulting in a time cost of approximately 750 yuan. During implementation, the cost calculation tool can integrate map data and real-time traffic information to ensure accurate values. For the initial cost values, it is determined whether the path cost exceeds a preset threshold. If it exceeds, the batch path is adjusted through a genetic algorithm to obtain an optimized path plan, and this process aims to reduce the excessive cost.

[0142] In one embodiment, assume the threshold is 80 yuan and the cost of the northern batch of 100 yuan exceeds the standard. The genetic algorithm can simulate multiple path combinations and iteratively select the shortest path. For example, some orders are merged into adjacent routes, and the adjusted path cost is reduced to 75 yuan.

[0143] It should be noted that the algorithm needs to balance the path length and vehicle loading, and dynamic adjustment can be achieved through scheduling software during implementation. According to the optimized path plan, the change trend of the time cost is obtained, and the genetic algorithm is used to iterate the schedule to determine the time sequence for efficient delivery, with the aim of optimizing the delivery rhythm.

[0144] Exemplarily, the adjusted path may shorten the delivery duration of some orders, such as from 15 minutes to 12 minutes. The genetic algorithm further optimizes the schedule and gives priority to arranging orders during peak periods to ensure the overall timeliness.

[0145] For example, the total time cost of the adjusted batch is reduced to 600 yuan. During implementation, the time sequence can be automatically generated by the scheduling system, and the efficiency is improved by combining the driver's shift schedule. Through the time sequence and the path plan, the route planning for each delivery batch is calculated to obtain the final route distribution data, and the core is to form an executable delivery plan.

[0146] In one possible implementation, the northern batch is planned into 3 routes, each covering 30 - 40 orders. The route distribution data includes the starting point, ending point, and passing points. During implementation, the navigation system can directly import the data, and the driver executes according to the plan to ensure that all orders are covered. The route distribution data is obtained, and it is determined whether the time cost meets the preset threshold. If not, the grouping method of task aggregation is adjusted, and the optimization iteration process is repeated, emphasizing dynamic error correction.

[0147] It can be understood that if the time cost still exceeds the 600 - yuan threshold, it can be regrouped. For example, the northern batch is split into two subgroups, and the paths and times are optimized respectively. During implementation, the scheduling center can judge in real time through the monitoring system and quickly iterate the grouping. According to the adjusted grouping method, the cost calculation method is used to re - determine the path cost and time cost to obtain the schedule and route planning for the efficiently delivered batches, ensuring the optimal plan.

[0148] For example, after regrouping, the cost of the northern batch drops to 70 yuan, the time cost is 550 yuan, and the schedule is accurate to the departure time of each order. During implementation, the system can generate visual reports for easy verification by dispatchers to ensure efficient delivery.

[0149] Step S107, update the real-time status of the cooperation between unmanned delivery technology and manual delivery through the schedule and route planning of efficient delivery batches, collect feedback data from the delivery execution process, and if the feedback data indicates resource idleness or task backlog, trigger the dynamic adaptation algorithm to readjust task decomposition and aggregation to obtain a new batch scheduling result.

[0150] Real-time collect feedback data during delivery execution through sensors and the log system to obtain the current status of task execution. If the feedback data exceeds the preset resource idleness threshold or task backlog threshold, activate the dynamic adaptation algorithm to determine the adjustment requirements. Process the feedback data using the dynamic adaptation algorithm to obtain a preliminary adjustment plan for task decomposition and task aggregation. Obtain a new batch scheduling arrangement based on the optimized task decomposition result. Generate corresponding route planning data according to the batch scheduling arrangement. After obtaining the route planning data, update the real-time status of unmanned delivery and manual delivery to obtain a collaborative execution plan. Adjust the delivery execution process through the collaborative execution plan to obtain new feedback data. Loop input

[0151] Specifically, the real-time collection of feedback data during delivery execution by sensors and the log system is the basis for building a dynamic delivery system.

[0152] Exemplarily, the sensors may include the positioning device, load sensor, and environmental perception device of the unmanned delivery vehicle, and the log system records the task completion time and abnormal events. Assume that a delivery center collects data once a minute to obtain information such as the current position, remaining power, and task completion rate of the delivery vehicle. These data reflect the real-time status of task execution. For example, if a vehicle stays in a certain area for too long, it may indicate traffic congestion or equipment failure.

[0153] In a possible implementation, if the feedback data exceeds the preset resource idleness threshold or task backlog threshold, an adjustment needs to be triggered.

[0154] For example, the resource idleness threshold is set to the vehicle's empty load rate exceeding 70%, and the task backlog threshold is set to more than 20 orders to be delivered in a certain area. Assume that the empty load rate of a delivery vehicle reaches 80%, or there are 25 orders piled up in a certain area, and the system will automatically mark it as abnormal and activate the dynamic adaptation algorithm. The core of this algorithm lies in quickly responding to abnormal states.

[0155] Preferably, by analyzing historical data and current feedback, determine whether tasks need to be reallocated or routes need to be adjusted.

[0156] Specifically, when the dynamic adaptation algorithm processes feedback data, it generates task decomposition and aggregation schemes based on task priorities and the real-time status of delivery vehicles.

[0157] For example, when there is a surge in orders in a certain area, the algorithm can split large orders into multiple small tasks and assign them to nearby idle vehicles; or aggregate small orders in adjacent areas into one batch and assign them to a single vehicle for execution. Suppose a vehicle was originally scheduled to deliver 10 orders. After algorithm adjustment, it is split into 5 orders and 5 orders, which are completed by two vehicles respectively, thus improving efficiency.

[0158] It should be noted that through the optimized task decomposition results, the system generates a new batch scheduling arrangement.

[0159] For example, after adjustment, the task volume of each vehicle is balanced, and the delivery time window is compressed from the original 2 hours to 1.5 hours. The scheduling arrangement also takes into account the vehicle's battery power and road conditions, and preferentially selects vehicles with sufficient battery power to perform long-distance tasks. Such an arrangement ensures the maximization of resource utilization.

[0160] In one embodiment, when generating route planning data according to the batch scheduling arrangement, the system comprehensively considers the real-time road conditions and the locations of orders.

[0161] For example, a vehicle needs to deliver 3 orders located at points A, B, and C respectively. The system selects the route from A to B and then to C based on the road conditions, avoiding congested areas. This kind of planning data directly affects the delivery efficiency.

[0162] Preferably, buffer time is reserved to cope with emergencies.

[0163] It can be understood that after obtaining the route planning data, the real-time status of unmanned delivery and manual delivery is updated to form a collaborative execution plan.

[0164] For example, unmanned vehicles are responsible for delivering on the main roads, and manual delivery workers take over the last mile. Suppose an unmanned vehicle runs out of power after completing 80% of the tasks. The system transfers the remaining tasks to the manual delivery worker to ensure the orders are delivered on time. This collaborative plan improves the flexibility of delivery.

[0165] For example, after adjusting the delivery process through the collaborative execution plan, the system will generate new feedback data and input it in a loop.

[0166] For example, after adjustment, the task backlog in a certain area drops from 25 orders to 10 orders, and the vehicle empty load rate drops from 80% to 50%. These data re-enter the system to form a closed-loop optimization. This continuous feedback loop ensures that the delivery system always operates efficiently and adapts to complex and changing environments.

[0167] In step S108, after obtaining the new batch scheduling result, the resource centralized management module synchronously updates the running status and location information of all distribution resources, generates the task decomposition and aggregation input data for the next cycle, and determines the continuous optimization ability of the system in the dynamic environment.

[0168] After obtaining the batch scheduling result, the management module synchronously updates the running status and location information to obtain the latest distribution data of the distribution resources. Extract the task decomposition elements from the latest distribution data, use the clustering algorithm to generate the aggregation data, and determine the basis for task allocation in the next cycle. Analyze the change trend of the dynamic environment through the aggregation data, and judge whether the current optimization ability of the system meets the requirements. If the change trend exceeds the preset threshold, adjust the running status through resource management to obtain the optimized resource distribution data. According to the optimized resource distribution data, generate the batch scheduling input for the next cycle and determine the priority sequence of task decomposition. Use the regression algorithm to predict the impact of the dynamic environment on the optimization ability and judge the adaptability of the system in the next cycle. Combine the priority sequence with the aggregation data through the management module to generate the final task allocation plan.

[0169] Exemplarily, after obtaining the batch scheduling result, the process of the management module synchronously updating the running status and location information can be achieved by real-time monitoring of the locations of distribution vehicles and personnel. Suppose in an urban distribution network, there are unmanned delivery vehicles and human delivery staff. The management module uses GPS and Internet of Things devices to obtain the current location, speed, and task completion status of each vehicle in real time.

[0170] For example, after an unmanned delivery vehicle completes a task in area A, its location is updated to standby at point A. After a human delivery staff completes a delivery in area B, the status is updated to be able to accept new tasks. The latest distribution data shows that there are 2 unmanned vehicles on standby in area A and 3 delivery staff are idle in area B. This real-time update method ensures the accuracy of the resource distribution data and provides a reliable basis for subsequent task allocation.

[0171] In a possible implementation, when extracting the task decomposition elements from the latest distribution data, the geographical location, urgency, and resource matching degree of the tasks can be concerned. Suppose there are currently 10 delivery orders distributed in areas C and D. The management module will extract the delivery time requirements and cargo types of each order.

[0172] For example, most of the orders in area C are fresh food and require delivery within 30 minutes; the orders in area D are ordinary packages and have more time. Based on these elements, the clustering algorithm clusters the orders in area C into one category and preferentially allocates unmanned vehicles to ensure speed. The orders in area D are assigned to human delivery staff to utilize their flexibility. This clustering method can effectively match the task and resource characteristics.

[0173] Specifically, after using a clustering algorithm to generate aggregated data, the process of determining the basis for task allocation in the next cycle can be achieved by analyzing the geographic density and resource distribution of orders. For example, if orders in Region C are concentrated in a certain business district, the algorithm will aggregate these orders into a single delivery batch and assign them to a single unmanned vehicle. Meanwhile, orders in Region D are more dispersed, so the algorithm will split them into multiple smaller batches and assign them to different delivery drivers. The generation of aggregated data clearly outlines the priority of task allocation, laying the foundation for scheduling in the next cycle.

[0174] It should be noted that when analyzing the changing trends of a dynamic environment, the system optimization capability can be judged through historical data and real-time feedback.

[0175] For example, during a peak delivery period, the volume of orders in Area C suddenly increased by 20%, exceeding the autonomous vehicle's carrying capacity. The management module, through trend analysis, determined that the system needed to temporarily reallocate resources. If the trend exceeded a preset threshold, such as a backlog of more than 10 orders, the system, through the resource management module, deployed a human delivery driver from Area B to support Area C. This optimized resource distribution data showed that resource constraints in Area C had been alleviated, ensuring delivery efficiency.

[0176] In one embodiment, determining the priority sequence is crucial when generating the next batch scheduling input. Suppose the system prioritizes fresh produce orders in Area C, followed by general orders in Area D, based on order urgency. The management module generates input data indicating that the autonomous vehicle prioritizes Area C, followed by human delivery personnel in Area D. This priority sequence ensures that critical tasks are completed first.

[0177] Preferably, when using regression algorithms to predict dynamic environmental impacts, analysis can be performed based on historical delivery data.

[0178] For example, the system analyzed delivery records from the past week and found that rainy days increased delivery times by 15%. Based on this, the regression algorithm predicted that if rain were to occur in the next cycle, the system would need to increase resource investment by 10%. This prediction method enhances the system's adaptability to external environments.

[0179] For example, the management module combines the priority sequence with aggregated data to generate a final task allocation plan, which can be displayed through a visual interface. For example, suppose an unmanned vehicle in area C is assigned five fresh produce orders, while three delivery drivers in area D each receive two standard orders. The allocation plan clearly defines each driver's route and time requirements, ensuring efficient task execution. This combined approach makes the allocation plan both scientific and actionable.

[0180] It can be understood that the implementation of each of the above-mentioned links together constitutes an efficient closed-loop distribution resource management. From real-time update to final allocation, each link is closely connected and the data flows smoothly. The advantage of this method lies in its ability to dynamically respond to changes in distribution requirements, ensure the maximization of resource utilization, and at the same time take into account the timeliness of tasks and customer satisfaction.

[0181] In addition, it should be noted that the various specific technical features described in the above specific implementation manners can be combined in any appropriate manner without contradiction. To avoid unnecessary repetition, the present invention will not separately describe various possible combination manners. In addition, any combination can be made between various different embodiments of the present invention as long as it does not violate the idea of the present invention, and it should also be regarded as the content disclosed by the present invention.

Claims

1. A multimodal collaborative distribution scheduling method, characterized in that The method includes: S101. Obtain the terrain data, traffic condition data, and real-time order information within the delivery area, identify the available resource scope of the collaborative operation of the unmanned delivery technology and manual delivery through the multi-modal delivery mode, and determine the initial state and location distribution of each resource in the current period; S102. Construct a dynamic environment model based on the terrain data and traffic condition data, extract the delivery demand characteristics from the real-time order information, and use the task decomposition technology to split the complex order into several sub-task units to obtain the decomposed sub-task set; S103. For the decomposed sub-task set, obtain the geographical location, time window, and resource demand attributes of each sub-task, match the terrain and traffic conditions through the dynamic adaptability algorithm, and judge whether the sub-task conflicts with the current environmental conditions. If so, adjust the priority and execution order of the sub-task to obtain the optimized sub-task sequence; S104. Extract the sub-task characteristics of adjacent areas and similar time windows from the optimized sub-task sequence, perform clustering analysis on the sub-tasks using the intelligent combination technology, determine the spatio-temporal correlation between the sub-tasks, and generate a preliminary sub-task aggregation grouping; S105. Obtain the total resource demand and delivery path constraints of the preliminary sub-task aggregation grouping, allocate the resource combination of the collaborative operation of the unmanned delivery technology and manual delivery through the resource centralized management module, and judge whether the resource demand exceeds the current capacity. If so, allocate and supplement from the standby resource pool to obtain the final resource allocation plan; S106. Calculate the path cost and time cost of each delivery batch according to the final resource allocation plan and the sub-task aggregation grouping, and use the genetic algorithm to optimize and iterate the batch path to obtain the schedule and route plan of the efficient delivery batch; S107. Update the real-time state of the collaborative operation of the unmanned delivery technology and manual delivery through the schedule and route plan of the efficient delivery batch, collect feedback data during the delivery execution process, and judge whether the feedback data indicates resource idleness or task backlog. If so, trigger the dynamic adaptability algorithm to re-adjust the task decomposition and aggregation to obtain a new batch scheduling result; S108. After obtaining the new batch scheduling result, synchronously update the operation state and location information of all delivery resources through the resource centralized management module, generate the input data for task decomposition and aggregation in the next cycle, and determine the continuous optimization ability of the system in the dynamic environment.

2. The method according to claim 1, wherein The S101 includes: Obtain the terrain data and traffic condition data within the delivery area, and through a preset geographic information system, process to obtain the distribution of the accessible paths within the area. Combine the real-time order information with the distribution of the accessible paths, and use the K-means algorithm for clustering analysis to obtain the order-intensive areas and the distribution of resource requirements. If the traffic condition data in the order-intensive areas exceeds the preset threshold, then identify the available resource range for unmanned delivery through multimodal analysis. Determine the initial state and location distribution of the unmanned delivery equipment according to the available resource range for unmanned delivery and the distribution of real-time orders. Obtain the collaborative resource data for manual delivery, and combine it with the demand distribution in the order-intensive areas to judge the initial state and location distribution of manual delivery. Through the superposition analysis of the location distribution and the initial state, obtain the collaborative resource allocation plan for unmanned delivery and manual delivery during the current period. For the collaborative resource allocation plan, use the Dijkstra algorithm to calculate the optimal path distribution of each resource from the initial state to the target order.

3. The method according to claim 1, wherein The S102 includes: Construct a dynamic model through the fusion of terrain data and traffic conditions to obtain an environmental feature description; Adopt real-time update technology to adjust the dynamic model to determine the changing trend of traffic conditions; Extract the delivery demand characteristics from the order information to obtain the demand priority ranking; Perform task decomposition on the demand priority ranking to obtain a preliminary sub-task division; If there are overlaps in the sub-task division, then adjust the sub-task set through data fusion technology to judge the optimized sub-task set; Divide the units according to the optimized sub-task set to obtain the final delivery task allocation; Verify the final delivery task allocation through a machine learning algorithm to determine the task execution sequence.

4. The method according to claim 1, wherein The S103 includes: Obtain the geographical location, time window, and resource demand attributes in the sub-task set, and through database query, obtain the terrain traffic and environmental condition data; Match the geographical location with the terrain traffic data through a dynamic adaptability algorithm to judge whether the sub-task conflicts with the environmental conditions, and obtain the conflict judgment result; If the conflict judgment result is true, then adjust the priority according to the time window to obtain the adjusted priority sequence; According to the adjusted priority sequence, rearrange the execution order to obtain a preliminarily sorted sub-task sequence; Analyze the preliminarily sorted sub-task sequence through the resource demand attributes, and use the greedy algorithm to optimize the resource allocation to obtain the sub-task sequence after resource optimization; Verify the matching consistency of the geographical location and the terrain traffic for the sub-task sequence after resource optimization to obtain the final optimized sequence; Through the final optimized sequence, update the execution order of the sub-task set to determine the task scheduling plan.

5. The method according to claim 1, characterized in that, The S104 includes: Obtain the task feature data through the sub-task sequence, and process it using feature extraction technology to obtain a feature set; Extract the adjacent area and time window information from the feature set to determine the distribution of sub-tasks with similar areas and similar times; For the data with similar areas and similar times, use intelligent combination technology for grouping to obtain a preliminary clustering result; Perform spatio-temporal correlation analysis on the preliminary clustering result to judge the correlation strength between sub-tasks and determine the optimized clustering grouping; Obtain the optimized clustering groups, and use the K-means algorithm to perform secondary clustering on the subtask sequence to obtain refined aggregation groups; Analyze the matching degree of task characteristics and spatio-temporal correlation through the refined aggregation groups, judge the grouping consistency, and obtain the final subtask aggregation groups; Extract task optimization information from the final subtask aggregation groups, and determine the execution order and resource allocation plan among subtasks.

6. The method according to claim 1, characterized in that The S105 includes: Obtain the resource demand quantity and distribution path constraints of the subtask aggregation groups through the task decomposition module, determine the preliminary demand distribution data, process the distribution path constraints using the path optimization algorithm, and obtain the optimized path allocation plan. Integrate the resource combinations of unmanned delivery technology and manual delivery collaboration through the resource centralized management module, and judge the feasibility of the allocation. If the demand exceeds the current capacity limit, analyze the excess part through the capacity assessment tool to obtain the available resource data of the backup resource pool. Allocate supplementary resources from the backup resource pool to the resource combination allocation to obtain the adjusted resource configuration result. Optimize the adjusted resource configuration result using the linear programming algorithm to determine the final allocation plan. Verify the execution status of the final allocation plan through the real-time monitoring module to obtain the dynamic adjustment data during the execution process.

7. The method according to claim 1, characterized in that, The S106 includes: Divide the subtask groups through the resource allocation plan, calculate the grouped results after task aggregation, and obtain the basic data of the delivery batches. Extract the path cost and time cost from the basic data of the delivery batches, and use the cost calculation method to determine the initial cost value of each batch. For the initial cost value, judge whether the path cost exceeds the preset threshold. If it exceeds, adjust the batch path through the genetic algorithm to obtain the optimized path plan. According to the optimized path plan, obtain the change trend of the time cost, and use the genetic algorithm to iterate the time schedule to determine the time series for efficient delivery. Calculate the route planning of each delivery batch through the time series and the path plan to obtain the final route distribution data. Obtain the route distribution data and judge whether the time cost meets the preset threshold. If it does not meet, adjust the grouping method of task aggregation and repeat the optimization iteration process. According to the adjusted grouping method, use the cost calculation method to re-determine the path cost and time cost to obtain the time schedule and route planning for efficient delivery batches.

8. The method according to claim 1, wherein The S107 includes: Real-time collect the feedback data during the delivery execution through the sensor and the log system to obtain the current state of task execution. If the feedback data exceeds the preset resource idle threshold or task accumulation threshold, activate the dynamic adaptability algorithm to determine the adjustment requirements. Process the feedback data using the dynamic adaptability algorithm to obtain the preliminary adjustment plan for task decomposition and task aggregation. Obtain the new batch scheduling arrangement according to the optimized task decomposition result. Generate the corresponding route planning data according to the batch scheduling arrangement. After obtaining the route planning data, update the real-time states of unmanned delivery and manual delivery to obtain the collaborative execution plan. Adjust the delivery execution process through the collaborative execution plan to obtain the new feedback data and loop input.

9. The method according to claim 1, characterized in that, The S108 includes: After obtaining the batch scheduling result, the management module synchronously updates the running status and location information to obtain the latest distribution data of the distribution resources; Extract the task decomposition elements from the latest distribution data, use the clustering algorithm to generate aggregated data, and determine the task allocation basis for the next cycle; Analyze the change trend of the dynamic environment through the aggregated data, and judge whether the current optimization ability of the system meets the requirements; If the change trend exceeds the preset threshold, adjust the running status through resource management to obtain the optimized resource distribution data; Generate the batch scheduling input for the next cycle according to the optimized resource distribution data, and determine the priority sequence of task decomposition; Use the regression algorithm to predict the impact of the dynamic environment on the optimization ability, and judge the adaptability of the system in the next cycle; Combine the priority sequence with the aggregated data through the management module to generate the final task allocation plan.

10. A multi-modal collaborative distribution scheduling system, characterized in that, Including: Multi-modal data acquisition module: Configured to acquire terrain data, traffic condition data and real-time order information in the distribution area, process through the geographic information system to obtain the distribution of passage paths, combine the K-means algorithm for clustering analysis of order-intensive areas and resource demand distribution, identify the available resources for unmanned distribution and the collaborative resource data for manual distribution, determine the initial state, location distribution and collaborative resource allocation plan of unmanned distribution equipment and manual distribution, and use the Dijkstra algorithm to calculate the optimal path distribution; Dynamic environment modeling module: Configured to fuse terrain data and traffic conditions to build a dynamic model, adjust the model through real-time update technology to determine the change trend of traffic conditions, extract distribution demand characteristics from order information and perform priority sorting, use task decomposition technology to split complex orders into sub-task sets, and optimize the sub-task division and verify the task execution sequence through data fusion technology; Task decomposition and optimization module: Configured to obtain the geographical location, time window and resource demand attributes of sub-tasks, match the terrain and traffic data through the dynamic adaptability algorithm to judge the conflict between tasks and environmental conditions, adjust the sub-task priority and execution order if there is a conflict, optimize the resource allocation in combination with the greedy algorithm, verify the matching consistency of the geographical location and terrain traffic, and generate the finally optimized sub-task sequence; Sub-task clustering and aggregation module: Configured to extract the adjacent area and time window characteristics of sub-tasks, use intelligent combination technology and K-means algorithm for clustering analysis, determine the spatio-temporal correlation between sub-tasks, and generate refined sub-task aggregation groups; Resource centralized management module: Configured to obtain the total resource demand and distribution path constraints of sub-task aggregation groups, integrate the resource combination of unmanned distribution technology and manual distribution collaboration, allocate and supplement from the standby resource pool if the resource demand exceeds the current capacity, optimize through the linear programming algorithm to obtain the final resource allocation plan, and monitor the execution status of the plan in real time; Path and time optimization module: Configured to calculate the path cost and time cost according to the resource allocation plan and sub-task aggregation groups, use the genetic algorithm to optimize and iterate the batch path and schedule, and generate the schedule and route plan for efficient distribution batches; Dynamic feedback adjustment module: Configured to collect distribution execution feedback data through sensors and the logging system. If resource idleness or task backlog is detected, it activates the dynamic adaptability algorithm to re-adjust task decomposition and aggregation, generates a new batch scheduling result, synchronously updates the operating status and location information of distribution resources, generates the input data for task decomposition and aggregation in the next cycle, and predicts the continuous optimization ability of the system in a dynamic environment through a regression algorithm.

Images