A stepwise hierarchical transport capacity scheduling and multi-objective optimization scheduling system and method

Abstract

The application discloses a kind of stepwise hierarchical transport capacity scheduling and multi-objective optimization scheduling system and method, belong to terminal logistics intelligent scheduling technical field, to solve the problem of existing scheduling logic solidification, transport capacity matching success rate is low, multi-objective is difficult to take into account, two-way business multiplexing degree is low.System includes whole link information synchronization unit and stepwise transport capacity scheduling unit, terminal equipment is divided into first level transfer transport capacity equipment and second terminal transport capacity equipment;Hierarchical transport capacity matching subunit executes hierarchical dynamic allocation, abnormal switching and combined link matching according to task priority, supports two-way business and return task merging and scheduling multiplexing;Multi-objective scheduling algorithm subunit constructs multi-objective integer programming model, solves and filters global optimal scheme using non-dominated sorting genetic algorithm with elitist strategy.The application can improve transport capacity matching efficiency, realize multi-objective dynamic balance, reduce equipment empty running rate, adapt to multi-scenario terminal unmanned distribution scheduling demand.

Description

[0001] A tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system and method This application is a divisional application of patent application No. 202611156467.4, filed on July 31, 2026, entitled "An unmanned delivery collaborative scheduling system, method and storage medium supporting two-way logistics". Technical Field

[0002] This invention relates to the field of intelligent scheduling technology for last-mile logistics, and in particular to a tiered hierarchical capacity scheduling and multi-objective optimization scheduling system and method, which can be applied to unmanned delivery and pickup scenarios for multiple categories of goods, such as express delivery and instant delivery. Background Technology

[0003] With the implementation of last-mile unmanned delivery technology, various heterogeneous devices such as shuttle delivery vehicles and delivery robots are gradually being deployed in communities, industrial parks, and other scenarios. Most existing unmanned delivery scheduling solutions are designed for a single type of equipment, with fixed scheduling logic, resulting in the following technical shortcomings in actual operation: First, the capacity matching mechanism is rigid, only supporting single-level single-time retrieval and scheduling, lacking hierarchical retry and cross-device automatic switching logic. During peak hours, capacity matching is prone to failure, while during off-peak hours, capacity is idle and wasted, resulting in poor adaptability. Second, the scheduling objective is singular, mostly focusing on the shortest timeliness, and cannot simultaneously take into account multiple dimensions such as operating costs and equipment load balancing, making it difficult to achieve dynamic balance in different business periods. Third, the two-way coordination capability is weak. Forward delivery and reverse pickup, as well as return and return tasks, need to be scheduled independently, making it impossible to reuse the same route capacity. The equipment empty running rate is high, and the overall operating efficiency is low. Summary of the Invention

[0004] To address the shortcomings of existing last-mile unmanned delivery scheduling schemes, such as rigid logic, low success rate of single-level capacity matching, inability to balance multiple objectives, and low reusability of bidirectional business capacity, this invention provides a tiered hierarchical capacity scheduling and multi-objective optimization scheduling system and method. Through a hierarchical dynamic matching mechanism and a multi-objective optimization algorithm, it achieves efficient capacity scheduling and bidirectional business reuse.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A tiered hierarchical capacity scheduling and multi-objective optimization scheduling system includes a full-link information synchronization unit and a tiered capacity scheduling unit deployed on a scheduling server. Each unit interacts with the end-terminal equipment via a wireless communication network. The terminal equipment includes primary transit equipment and secondary terminal equipment, which are used to perform express delivery, distribution and collection operations, and to exchange equipment status and task data in real time. The full-link information synchronization unit is used to build a globally unique business task database, realize the real-time synchronization of task information and equipment status among all devices in the full-link, and provide data support for capacity scheduling. The tiered capacity scheduling unit includes a hierarchical capacity matching subunit and a multi-objective scheduling algorithm subunit; The hierarchical capacity matching subunit is used to perform hierarchical dynamic capacity allocation and abnormal retry switching according to task priority. It supports two-way business of forward delivery and reverse pickup, and supports merging return and return tasks with normal business tasks of forward delivery and reverse pickup on the same route to realize capacity reuse. The multi-objective scheduling algorithm subunit is used to construct a constrained multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest transportation cost, and balanced equipment load. It is solved by a non-dominated sorting genetic algorithm with an elite strategy, and after outputting the Pareto optimal solution set, the globally optimal scheduling scheme is selected based on preset weight coefficients.

[0006] Furthermore, the hierarchical matching logic of the hierarchical capacity matching subunit is specifically as follows: A preset configurable search failure threshold is provided. Non-urgent priority tasks prioritize searching for and matching the corresponding level of first- or second-level transportation equipment. If the number of search failures exceeds the search failure threshold, the search will automatically switch to matching the next level of transportation equipment. Priority tasks such as scheduled delivery and scheduled pickup can continue to search for and match adjacent levels of transportation equipment based on the business scenario after successfully matching the corresponding level of transportation equipment, forming a combined delivery link. Urgent priority tasks are directly matched with second-level last-mile transportation equipment. If the search and matching of secondary end-point transportation equipment still fails and the number of failures exceeds the preset search failure threshold, the user interaction options are triggered, supporting cancellation of the task to trigger return of the item or continued retry.

[0007] Furthermore, the priorities for forward delivery, from highest to lowest, are: emergency delivery priority, forward return and return priority, and scheduled delivery priority. Prioritize the search for and match primary transit equipment for scheduled delivery tasks. If a match is found, continue to search for and match secondary terminal equipment to form a transit + terminal combined delivery link. If the number of failed searches exceeds the search failure threshold, switch to searching for and matching secondary terminal equipment to perform direct delivery.

[0008] Furthermore, the priority of reverse pickup, from high to low, is as follows: emergency pickup priority, reverse return and return priority, and scheduled pickup priority; Priority tasks for scheduled pickups prioritize searching for and matching secondary last-mile delivery equipment. If a successful search and match is found, the last-mile pickup process is marked as successfully allocated, and the search and matching of primary transit equipment can continue. If a primary transportation equipment search and match is successful, a last-mile pickup + transit transportation combined link is formed. If the number of primary transportation equipment search failures exceeds the search failure threshold, the matched secondary last-mile delivery equipment will perform the entire pickup task.

[0009] Furthermore, return and return tasks are matched with corresponding level of transport capacity equipment according to the current transportation node of the express; among them, forward return and return tasks are merged with reverse pickup tasks along the same route, and reverse return and return tasks are merged with forward delivery tasks along the same route, sharing equipment load quotas.

[0010] Furthermore, the multi-objective scheduling algorithm subunit is configured to solve for minimizing the objective function: F = ω1×T + ω2×C + ω3×L In the formula, T is the total delivery time of all tasks, C is the total cost of transportation capacity, L is the load imbalance of all devices, and ω1, ω2, and ω3 are the weight coefficients corresponding to timeliness, cost, and load balancing, respectively. The weight coefficients can be dynamically adjusted according to the business scenario.

[0011] Furthermore, the constraints of the multi-objective integer programming model include equipment endurance constraints, load capacity constraints, time constraints, placement and handover time constraints, restricted area constraints, total travel constraints, and building access time constraints.

[0012] Furthermore, the non-dominated sorting genetic algorithm with elite strategy uses integer encoding to encode the task-device allocation relationship, and the constraint conditions are handled by the penalty function method; the elite retention ratio, crossover probability, mutation probability, and iteration termination condition of the algorithm are all preset configuration parameters, and each parameter is configured with a corresponding preset value range, which can be dynamically adjusted within the corresponding value range based on the scale and timeliness requirements of the business scenario.

[0013] Furthermore, the preset value ranges for each parameter are as follows: the elite retention rate is 15%~25%, the crossover probability is 0.7~0.9, the mutation probability is 0.05~0.15, and the iteration termination condition is that the optimization amplitude of the Pareto front for 15~30 consecutive generations is less than the preset convergence threshold.

[0014] Furthermore, the preferred values ​​for the preset configuration parameters are: an elite retention rate of 20%, a crossover probability of 0.8, a mutation probability of 0.1, and an iteration termination condition where the optimization magnitude of the Pareto front for 20 consecutive generations is less than a preset convergence threshold. The preset convergence threshold is used to determine the algorithm's convergence state and can be configured according to computational cost and solution accuracy requirements.

[0015] Furthermore, the primary transit equipment is a shuttle delivery vehicle, the secondary terminal equipment is a delivery robot, and the terminal equipment also includes a smart parcel locker as a storage and connection device.

[0016] A tiered, hierarchical capacity scheduling and multi-objective optimization scheduling method, characterized by being executed by a scheduling server, includes the following steps: Step 1: End-to-end data synchronization. The scheduling server builds a globally unique business task database and interacts with the terminal devices, which include primary transit capacity equipment, secondary terminal capacity equipment, and storage and connection equipment, in real time to exchange task information and equipment status data, providing consistent data support for capacity scheduling. Step 2: Task priority classification. The scheduling server classifies forward delivery tasks and reverse pickup tasks into priority levels according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and bidirectional reuse scheduling. The scheduling server merges returned shipment tasks with normal business tasks along the same route and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level of capacity equipment. After the number of failed searches exceeds the preset failure threshold, the server automatically switches to searching for and matching the next level of capacity equipment. Priority tasks such as scheduled delivery and scheduled pickup can continue to search for and match adjacent level capacity equipment to form a combined delivery link after a successful matching of the corresponding level of capacity. Urgent priority tasks directly search for and match secondary last-mile capacity equipment. Step 4: Multi-objective global optimization solution. After the scheduling server locks the candidate capacity pool, it constructs a multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. It uses a non-dominated sorting genetic algorithm with elitist strategy to solve the problem. After outputting the Pareto optimal solution set, it selects the globally optimal scheduling scheme based on the preset weight coefficients. Step 5: Issuance of scheduling instructions. The scheduling server issues task execution instructions to the corresponding end terminal devices according to the globally optimal scheduling scheme.

[0017] A tiered, hierarchical capacity scheduling and multi-objective optimization scheduling method, implemented based on the above system, includes the following steps: Step 1: End-to-end data synchronization. The end-to-end information synchronization unit constructs a globally unique business task database to achieve real-time synchronization of task information and equipment status between end-to-end terminal equipment (including primary transit capacity equipment, secondary terminal capacity equipment, and storage connection equipment) and the tiered capacity scheduling unit. Step 2: Task priority determination. The tiered capacity scheduling unit divides the priority levels of forward delivery tasks and reverse collection tasks according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and bidirectional capacity reuse scheduling. The tiered capacity matching sub-unit merges returned shipment tasks with normal business tasks along the same route for scheduling and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level capacity equipment. After the number of failed searches exceeds the preset failure threshold, the search and matching will automatically switch to the next level of capacity equipment. After successful matching of the corresponding level of capacity for scheduled delivery and scheduled pickup, priority tasks can continue to search for and match adjacent level capacity equipment according to the business scenario to form a combined delivery link. Urgent priority tasks directly search for and match secondary last-mile capacity equipment. Step 4: Multi-objective optimization solution. After the multi-objective scheduling algorithm sub-unit locks the candidate capacity pool, it constructs a multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. It is solved by a non-dominated sorting genetic algorithm with an elite strategy, outputs the Pareto optimal solution set, and selects the globally optimal scheduling scheme. Step 5: Task execution and status synchronization. Task instructions are sent to each end terminal device, and each end terminal device performs the corresponding transfer, delivery or pickup operation and transmits task status data back in real time.

[0018] Furthermore, in step 3 of any of the above methods: The priority of forward delivery, from high to low, is: emergency delivery priority, forward return and return priority, and scheduled delivery priority. Scheduled delivery priority tasks are first matched with primary transit capacity equipment. After a successful search and match, they continue to be matched with secondary terminal capacity equipment to form a transit + terminal combined delivery link. If the number of failed searches exceeds the preset search failure threshold, the search and matching will switch to secondary terminal capacity equipment to perform direct delivery. The priority of reverse pickup, from high to low, is as follows: emergency pickup priority, reverse return and return priority, and scheduled pickup priority. Scheduled pickup priority tasks are matched with secondary last-mile delivery equipment first. If the search is successful, the last-mile pickup link is marked as successfully allocated, and the search can continue to match primary transit delivery equipment. If the primary delivery equipment search is successful, a last-mile pickup + transit transportation combined link is formed. If the number of primary delivery equipment search failures exceeds the preset search failure threshold, the matched secondary last-mile delivery equipment will perform the entire pickup task.

[0019] Furthermore, in step 4 of any of the above methods, the model constraints include equipment endurance, load capacity, timeliness, placement and handover time, restricted areas, total travel distance, and building access time; After the algorithm outputs the Pareto optimal solution set, it uses weighted selection based on the weight coefficients corresponding to the business scenario to select the globally optimal scheduling scheme.

[0020] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above methods.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Improve the success rate of capacity matching and scenario adaptability. Through a tiered retry and cross-level automatic switching mechanism, non-urgent tasks prioritize the use of low-cost, high-capacity equipment to complete batch transfers, while urgent tasks are directly matched with high-efficiency last-mile delivery capacity, balancing cost and timeliness, and significantly reducing the probability of capacity matching failure.

[0022] 2. Achieve global dynamic balance across multiple objectives. Based on a multi-objective integer programming model and a non-dominated sorting genetic algorithm with an elitist strategy, it simultaneously optimizes three major objectives: timeliness, cost, and load balancing. The weight coefficients can be dynamically adjusted according to business scenarios, prioritizing cost reduction during off-peak periods and timeliness during peak periods to adapt to different operational needs.

[0023] 3. Reduce equipment empty running rate and improve capacity utilization. Support the merging of forward delivery, reverse pickup and return tasks, and capacity reuse. Tasks on the same route share equipment load quotas, effectively reducing ineffective driving and lowering the overall operating cost at the last mile. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the tiered hierarchical transportation capacity scheduling system architecture of the present invention; Figure 2 This is a schematic diagram of the forward delivery hierarchical capacity matching process of the present invention; Figure 3 This is a schematic diagram of the reverse collection and tiered capacity matching process of the present invention; Figure 4 This is a schematic diagram of the solution process for the multi-objective optimization scheduling algorithm of the present invention. Detailed Implementation

[0025] The principles and features of the present invention are described below with reference to the accompanying drawings. The embodiments described are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0026] The primary transfer equipment is a shuttle delivery vehicle, used for the transfer of bulk express parcels between multiple locations.

[0027] The secondary last-mile delivery equipment is a delivery robot, used to perform door-to-door delivery and pickup tasks.

[0028] The storage and connection equipment is a smart parcel locker, used for the temporary storage and sorting of parcels and for connection with primary transit equipment and secondary terminal transportation equipment.

[0029] The parameter N is the preset search failure threshold, which is a preset positive integer. The recommended value is 5, and the value range is 2~10.

[0030] The parameter S is the search interval, with a recommended value of 10 seconds and a range of 5 to 20 seconds.

[0031] The parameter AS is the waiting time for user interaction options; a value of 60 seconds is recommended.

[0032] Forward return and return: This refers to the service where items awaiting delivery are returned to their original collection point during the forward delivery process.

[0033] Reverse return of packages: The service where items to be delivered are returned to the original sender during the reverse pickup process.

[0034] The capacity levels are defined from high to low as Level 1 transit capacity equipment and Level 2 terminal capacity equipment.

[0035] like Figures 1 to 3 As shown, a tiered hierarchical capacity scheduling and multi-objective optimization scheduling system is deployed in the scheduling central layer of an unmanned delivery collaborative scheduling system. It includes a full-link information synchronization unit and a tiered capacity scheduling unit. The tiered capacity scheduling unit is further divided into a hierarchical capacity matching subunit and a multi-objective scheduling algorithm subunit. The system interacts bidirectionally with the primary intermediate transport capacity equipment, the secondary terminal transport capacity equipment, and the storage and connection equipment through a wireless communication network.

[0036] The end-to-end information synchronization unit constructs a globally unique business task database, performs spatiotemporal benchmark alignment and unified encoding on all task data and device status data, provides a consistent data foundation for scheduling operations, and eliminates scheduling deviations caused by inconsistencies in data from multiple devices.

[0037] The hierarchical capacity matching subunit executes hierarchical dynamic allocation logic: The system presets a failure threshold N and a search interval S, and initializes the number of searches to 0. Non-urgent priority tasks are matched with the corresponding level of capacity first. If the matching failure exceeds the threshold, the capacity level can be switched down. After a successful match for scheduled delivery or scheduled pickup priority tasks, the system can continue to search and match adjacent levels of capacity according to the business scenario to form a complete delivery link.

[0038] For forward delivery missions: Emergency delivery priority: Skip the first-level transit capacity search and proceed directly to the second-level last-mile delivery capacity equipment search and matching; Scheduled delivery priority: Prioritize searching for and matching first-level transit equipment that meets the requirements in terms of load capacity, storage space, and driving route; if a search and match is successful, the transit link is marked as successfully allocated, the search count is reset, and the search continues to match second-level last-mile delivery equipment to perform last-mile delivery, forming a transit + last-mile combined delivery link; if the search for first-level equipment fails and the number of searches exceeds N, the search count is reset, and the system automatically switches to second-level last-mile delivery equipment to perform direct delivery throughout the entire process. Priority for forward return shipments: Match the corresponding level of transportation capacity according to the current transportation node of the shipment, and merge the shipment with reverse pickup tasks along the same route in the order scheduling.

[0039] For reverse pickup tasks: Emergency pickup priority: Directly search and match the nearest available secondary last-mile delivery equipment within the sender's address range; Priority for scheduled pickup: Prioritize searching for and matching secondary last-mile delivery equipment within the sender's address range that is along the route and meets the load capacity requirements; if the search is successful, the allocation is marked as successful, the search count is reset, and the search can continue to match primary transit equipment; if the search is successful, the transit link is marked as successfully allocated, forming a last-mile pickup + transit transportation combined link with the searched and matched secondary last-mile delivery equipment; if the primary equipment search fails and the number of searches exceeds N, the search count is reset, and the successfully matched secondary last-mile delivery equipment will perform the entire pickup task; Reverse return shipment priority: Match the corresponding level of transportation capacity according to the current transportation node of the shipment, and merge the shipment with the forward delivery task of the same route in the order scheduling.

[0040] If the search for secondary end-point transportation equipment still fails and exceeds the search failure threshold N, the user interaction options are triggered. The user can choose to cancel the task and trigger a return, or continue to retry; if the AS does not operate within the timeout period, the search will be re-executed.

[0041] The multi-objective scheduling algorithm subunit performs global optimization calculations: After identifying the candidate capacity pool, a multi-objective integer programming model is constructed to minimize the objective function: F = ω1×T + ω2×C + ω3×L Where T is the total delivery time for all tasks, C is the total transportation cost, and L is the load imbalance of all equipment; ω1, ω2, and ω3 are the weight coefficients corresponding to the three optimization objectives of timeliness, cost, and load balancing, respectively. The weight coefficients are configured as ω1=0.4, ω2=0.35, and ω3=0.25 during the off-peak period and automatically switched to ω1=0.6, ω2=0.2, and ω3=0.2 during the peak period.

[0042] The model constraints include: equipment endurance constraints, load and space load capacity constraints, task time constraints, equipment placement and handover time constraints, restricted area constraints, total travel constraints, and building passage time constraints.

[0043] The solution employs a non-dominated sorting genetic algorithm (NSGA-II) with an elite strategy: the task-device allocation relationship is encoded using integer encoding, and the constraint conditions are handled using a penalty function method; the elite retention ratio, crossover probability, mutation probability, and iteration termination condition of the algorithm are all preset configuration parameters, and each parameter is configured with a corresponding preset value range, which can be dynamically adjusted within the corresponding range according to the scale of the business scenario, time requirements, and number of devices.

[0044] The preset value ranges for each parameter are as follows: elite retention rate of 15%~25%, crossover probability of 0.7~0.9, mutation probability of 0.05~0.15, and iteration termination condition of no significant optimization in the Pareto front for 15~30 consecutive generations. "No significant optimization" means that the change in the objective function value of the optimal solution in the Pareto front is less than a preset convergence threshold, which can be configured according to computational cost and solution accuracy requirements. The above value ranges include endpoint values ​​and are adapted to the last-mile logistics capacity scheduling scenario, balancing algorithm convergence speed, solution accuracy, and computational cost. Those skilled in the art can flexibly configure specific parameter values ​​within the above ranges according to the actual deployed equipment scale and task volume, all of which fall within the protection scope of this invention.

[0045] The above value range is an engineering optimization range for last-mile logistics capacity scheduling scenarios. If the elite retention ratio is below 15%, it is easy to cause the loss of excellent traits in the population and a decrease in convergence speed. If it is above 25%, it is easy to cause premature maturity of the population and get trapped in local optima. The range settings of crossover probability, mutation probability and iteration termination generation take into account both solution accuracy and computational efficiency, and can be flexibly adjusted according to the actual deployment equipment scale and task volume.

[0046] In the preferred configuration of this embodiment, the elite retention ratio is 20%, the crossover probability is 0.8, the mutation probability is 0.1, and the iteration termination condition is that there is no significant optimization in the Pareto front for 20 consecutive generations. After outputting the Pareto optimal solution set, the globally optimal scheduling scheme is selected based on the weight coefficient of the current scenario.

[0047] like Figure 4 As shown, the algorithm solution process is as follows: population initialization → non-dominated sorting → crowding calculation → selection crossover operation → selection mutation operation → generation of offspring population → merging parent and child populations → non-dominated sorting and crowding calculation again → elite retention to generate a new generation population → iteration termination judgment → output Pareto optimal solution set → weighted selection of optimal solution → output global optimal scheduling solution.

[0048] The steps of the scheduling server in executing the tiered capacity scheduling and multi-objective optimization scheduling method are as follows: Step 1: End-to-end data synchronization. The scheduling server builds a globally unique business task database and interacts with shuttle delivery vehicles, delivery robots, and smart parcel lockers in real time to exchange task information and equipment status data, providing consistent data support for capacity scheduling. Step 2: Task priority classification. The scheduling server classifies forward delivery tasks and reverse pickup tasks into priority levels according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and two-way reuse scheduling. The scheduling server merges returned shipment tasks with normal business tasks along the same route and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level of capacity equipment. If the search failure exceeds the preset search failure threshold, it will automatically switch to searching for and matching the next level of capacity equipment. Priority tasks such as scheduled delivery and scheduled pickup can continue to search for and match adjacent level capacity equipment to form a combined delivery link after the corresponding level of capacity matching is successful. Urgent priority tasks directly search for and match the second-level last-mile delivery equipment, the delivery robot. Step 4: Multi-objective global optimization solution. After the scheduling server locks the candidate capacity pool, it constructs a multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. It uses a non-dominated sorting genetic algorithm with elitist strategy to solve the problem. After outputting the Pareto optimal solution set, it selects the globally optimal scheduling scheme based on the preset weight coefficients. Step 5: Issuance of scheduling instructions. The scheduling server issues task execution instructions to the corresponding end terminal devices according to the globally optimal scheduling scheme.

[0049] Corresponding to the above system architecture, the execution steps of the tiered hierarchical capacity scheduling and multi-objective optimization scheduling method of the present invention are as follows: Step 1: End-to-end data synchronization. Build a globally unique business task database to achieve real-time synchronization of task information and equipment status among the end-to-end equipment shuttle delivery vehicles, delivery robots, smart parcel lockers, and tiered capacity scheduling units. Step 2: Task priority determination. The tiered capacity scheduling unit divides the priority levels of forward delivery tasks and reverse collection tasks according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and bidirectional reuse scheduling. The tiered capacity matching sub-unit merges returned shipment tasks with normal business tasks along the same route for scheduling and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level of capacity equipment. If the search fails and exceeds a preset threshold, it automatically switches to searching for and matching the next level of capacity equipment. After successful matching of the corresponding level of capacity, scheduled delivery and scheduled pickup priority tasks can continue to search for and match adjacent level capacity equipment according to the business scenario to form a combined delivery link. Urgent priority tasks directly search for and match the second-level last-mile delivery equipment, the delivery robot. Step 4: Multi-objective optimization solution. After the multi-objective scheduling algorithm sub-unit locks the candidate capacity pool, a multi-objective integer programming model is constructed with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. The model constraints include equipment endurance, load capacity, timeliness, placement and handover time, restricted areas, total journey, and building passage time. A non-dominated sorting genetic algorithm with an elite strategy is used to solve the problem, outputting a Pareto optimal solution set and selecting the globally optimal scheduling scheme. Step 5: Task execution and status synchronization. Task instructions are sent to each end terminal device, and each end terminal device performs the corresponding transfer, delivery or pickup operation and transmits task status data back in real time.

[0050] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the aforementioned tiered hierarchical capacity scheduling and multi-objective optimization scheduling method. The storage medium includes, but is not limited to, various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks. This storage medium can be deployed in embedded storage units of scheduling hub servers, edge node controllers, and end-device devices.

[0051] Example 1: Scheduling and Execution in a Three-Tier Architecture Scenario This embodiment applies to a three-tiered, fully configured deployment scenario consisting of smart parcel lockers, shuttle delivery vehicles, and delivery robots. The specific scheduling and execution steps are as follows: 1. Data base initialization: The end-to-end information synchronization unit accesses business system order data, real-time equipment status data, and environmental map data, completes data cleaning and spatiotemporal alignment, and builds a global task database.

[0052] 2. Priority Determination: Based on the timeliness requirements of express delivery, forward delivery tasks are divided into three levels: emergency delivery, forward return and return, and scheduled delivery. Reverse pickup tasks are also divided into three levels: emergency pickup, reverse return and return, and scheduled pickup. They are then sorted and dispatched in descending order of priority.

[0053] 3. Tiered Capacity Matching: Scheduled delivery tasks are prioritized for shuttle delivery vehicles to complete batch transfers between smart parcel lockers and multiple buildings. If the retrieval fails more than 5 times, the system automatically switches to delivery robots for direct delivery. Urgent tasks are directly matched with delivery robots for point-to-point delivery from smart parcel lockers to recipient addresses. Forward return and return tasks are matched with on-route capacity and merged with reverse pickup tasks in the order scheduling. Reverse return and return tasks are matched with on-route capacity and merged with forward delivery tasks in the order scheduling. Scheduled pickup tasks are prioritized for door-to-door pickup by delivery robots. After successful pickup, the system continues to search for shuttle delivery vehicles for batch transfer. If the transfer retrieval fails, the delivery robot will complete the entire pickup process.

[0054] 4. Multi-objective global optimization: After locking in candidate capacity, substitute the objective function and seven types of constraints, run the NSGA-II algorithm to iteratively solve the problem, output the optimal scheduling scheme for the current time period, and allocate tasks and travel routes.

[0055] 5. Task execution and status synchronization: The status of the equipment is synchronized in real time during the execution of the task. When equipment failure, abnormal road conditions or other scenarios occur, dynamic rescheduling is triggered to rerun the hierarchical matching and multi-objective solution process and update the scheduling scheme.

[0056] Example 2: Scheduling and Execution in a Two-Level Simplified Architecture Scenario This embodiment omits the shuttle delivery vehicle and consists of a two-tier architecture of smart parcel lockers and delivery robots, making it suitable for high-rise scenarios such as office buildings.

[0057] Scheduling logic adaptation and adjustment: All delivery, pickup, and return tasks are directly matched with the secondary last-mile delivery robot; the hierarchical matching logic is simplified to hierarchical retry within the same level of equipment; the multi-objective optimization model adds elevator travel time constraints, merges tasks by floor, and reduces elevator round trips; during the morning peak period, the timeliness weight is increased to prioritize the delivery of urgent items.

[0058] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A tiered, hierarchical capacity scheduling and multi-objective optimization scheduling system, characterized in that, It includes a full-link information synchronization unit and a tiered capacity scheduling unit deployed on the scheduling server. Each unit and the terminal equipment interact with each other through a wireless communication network. The terminal equipment includes primary transit equipment and secondary terminal equipment, which are used to perform express delivery, distribution and collection operations, and to exchange equipment status and task data in real time. The full-link information synchronization unit is used to build a globally unique business task database, realize the real-time synchronization of task information and equipment status among all devices in the full-link, and provide data support for capacity scheduling. The tiered capacity scheduling unit includes a hierarchical capacity matching subunit and a multi-objective scheduling algorithm subunit; The hierarchical capacity matching subunit is used to perform hierarchical dynamic capacity allocation and abnormal retry switching according to task priority. It supports two-way business of forward delivery and reverse pickup, and supports merging return and return tasks with normal business tasks of forward delivery and reverse pickup on the same route to realize capacity reuse. The multi-objective scheduling algorithm subunit is used to construct a constrained multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest transportation cost, and balanced equipment load. It is solved by a non-dominated sorting genetic algorithm with an elite strategy, and after outputting the Pareto optimal solution set, the globally optimal scheduling scheme is selected based on preset weight coefficients.

2. The tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system according to claim 1, characterized in that, The hierarchical matching logic of the hierarchical capacity matching subunit is as follows: A preset configurable search failure threshold is provided. Non-urgent priority tasks prioritize searching for and matching with first- or second-level transportation equipment of the corresponding level. After the number of search failures exceeds the search failure threshold, the search will automatically switch to matching with the next level of transportation equipment. Priority tasks such as scheduled delivery and scheduled pickup can continue to search for and match with adjacent levels of transportation equipment after a successful match with the corresponding level of transportation equipment, forming a combined delivery link. Urgent priority tasks are directly matched with second-level last-mile transportation equipment. If the search and matching of secondary end-point transportation equipment still fails and the number of failures exceeds the preset search failure threshold, the user interaction options are triggered, supporting cancellation of the task to trigger return of the item or continued retry.

3. The tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system according to claim 2, characterized in that, The priorities for forward delivery, from highest to lowest, are: emergency delivery priority, forward return and return priority, and scheduled delivery priority. Prioritize the search for and match primary transit equipment for scheduled delivery tasks. If a match is found, the search can continue to match secondary terminal equipment to form a transit + terminal combined delivery link. If the number of failed searches exceeds the search failure threshold, the search will switch to matching secondary terminal equipment for direct delivery.

4. The tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system according to claim 2, characterized in that, The priority of reverse pickup, from highest to lowest, is: emergency pickup priority, reverse return and return priority, and scheduled pickup priority. Priority tasks for scheduled pickups prioritize searching for and matching secondary last-mile delivery equipment. If a successful search and match is found, the last-mile pickup process is marked as successfully allocated, and the search and matching of primary transit equipment can continue. If a primary transportation equipment search and match is successful, a last-mile pickup + transit transportation combined link is formed. If the number of primary transportation equipment search failures exceeds the search failure threshold, the matched secondary last-mile delivery equipment will perform the entire pickup task.

5. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling system according to claim 2, characterized in that, Return shipment return tasks are matched with the corresponding level of transportation capacity equipment according to the current transportation node of the shipment; among them, forward return shipment return tasks are merged with reverse pickup tasks along the same route, and reverse return shipment return tasks are merged with forward delivery tasks along the same route, sharing equipment load quotas.

6. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling system according to claim 1, characterized in that, The multi-objective scheduling algorithm subunit is configured to solve for minimizing the objective function: F = ω1×T + ω2×C + ω3×L In the formula, T is the total delivery time of all tasks, C is the total cost of transportation capacity, L is the load imbalance of all devices, and ω1, ω2, and ω3 are the weight coefficients corresponding to timeliness, cost, and load balancing, respectively. The weight coefficients can be dynamically adjusted according to the business scenario.

7. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling system according to claim 6, characterized in that, The constraints of the multi-objective integer programming model include equipment endurance constraints, load capacity constraints, time constraints, placement and handover time constraints, restricted area constraints, total travel constraints, and building access time constraints.

8. The tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system according to claim 6, characterized in that, The non-dominated sorting genetic algorithm with elite strategy uses integer encoding to encode the task-device allocation relationship, and the constraint conditions are handled by the penalty function method. The elite retention ratio, crossover probability, mutation probability, and iteration termination condition of the algorithm are all preset configuration parameters, and each parameter is configured with a corresponding preset value range. The configuration can be dynamically adjusted within the corresponding value range based on the scale and timeliness requirements of the business scenario.

9. The tiered hierarchical transportation capacity scheduling and multi-objective optimization scheduling system according to claim 8, characterized in that, The preset value ranges for each parameter are as follows: elite retention ratio is 15% to 25%, crossover probability is 0.7 to 0.9, mutation probability is 0.05 to 0.15, and the iteration termination condition is that the optimization amplitude of the Pareto front for 15 to 30 consecutive generations is less than the preset convergence threshold.

10. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling system according to claim 9, characterized in that, The preferred values ​​for the preset configuration parameters are: an elite retention rate of 20%, a crossover probability of 0.8, a mutation probability of 0.1, and an iteration termination condition where the optimization magnitude of the Pareto front for 20 consecutive generations is less than the preset convergence threshold.

11. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling system according to claim 1, characterized in that, The primary transit equipment is a shuttle delivery vehicle, the secondary terminal equipment is a delivery robot, and the terminal equipment also includes a smart parcel locker as a storage and connection device.

12. A tiered, hierarchical capacity scheduling and multi-objective optimization scheduling method, characterized in that, Executed by the scheduling server, the process includes the following steps: Step 1: End-to-end data synchronization. The scheduling server builds a globally unique business task database and interacts with the terminal devices, which include primary transit capacity equipment, secondary terminal capacity equipment, and storage and connection equipment, in real time to exchange task information and equipment status data, providing consistent data support for capacity scheduling. Step 2: Task priority classification. The scheduling server classifies forward delivery tasks and reverse pickup tasks into priority levels according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and bidirectional reuse scheduling. The scheduling server merges returned shipment tasks with normal business tasks along the same route and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level of capacity equipment. After the number of failed searches exceeds the preset failure threshold, the server automatically switches to searching for and matching the next level of capacity equipment. Priority tasks such as scheduled delivery and scheduled pickup can continue to search for and match adjacent level capacity equipment to form a combined delivery link after a successful matching of the corresponding level of capacity. Urgent priority tasks directly search for and match secondary last-mile capacity equipment. Step 4: Multi-objective global optimization solution. After the scheduling server locks the candidate capacity pool, it constructs a multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. It uses a non-dominated sorting genetic algorithm with elitist strategy to solve the problem. After outputting the Pareto optimal solution set, it selects the globally optimal scheduling scheme based on the preset weight coefficients. Step 5: Issuance of scheduling instructions. The scheduling server issues task execution instructions to the corresponding end terminal devices according to the globally optimal scheduling scheme.

13. A tiered, hierarchical capacity scheduling and multi-objective optimization scheduling method, characterized in that, The system implementation based on any one of claims 1 to 11 includes the following steps: Step 1: End-to-end data synchronization. The end-to-end information synchronization unit constructs a globally unique business task database to achieve real-time synchronization of task information and equipment status between end-to-end terminal equipment (including primary transit capacity equipment, secondary terminal capacity equipment, and storage connection equipment) and the tiered capacity scheduling unit. Step 2: Task priority determination. The tiered capacity scheduling unit divides the priority levels of forward delivery tasks and reverse collection tasks according to the type of express delivery business and time requirements. Step 3: Tiered capacity matching and bidirectional capacity reuse scheduling. The tiered capacity matching sub-unit merges returned shipment tasks with normal business tasks along the same route for scheduling and shares equipment load quotas. Non-urgent priority tasks prioritize searching for and matching corresponding level capacity equipment. After the number of failed searches exceeds the preset failure threshold, the search and matching will automatically switch to the next level of capacity equipment. After successful matching of the corresponding level of capacity for scheduled delivery and scheduled pickup, priority tasks can continue to search for and match adjacent level capacity equipment according to the business scenario to form a combined delivery link. Urgent priority tasks directly search for and match secondary last-mile capacity equipment. Step 4: Multi-objective optimization solution. After the multi-objective scheduling algorithm sub-unit locks the candidate capacity pool, it constructs a multi-objective integer programming model with the optimization objectives of optimal timeliness, lowest capacity cost, and balanced equipment load. It is solved by a non-dominated sorting genetic algorithm with an elite strategy, outputs the Pareto optimal solution set, and selects the globally optimal scheduling scheme. Step 5: Task execution and status synchronization. Task instructions are sent to each end terminal device, and each end terminal device performs the corresponding transfer, delivery or pickup operation and transmits task status data back in real time.

14. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling method according to claim 12 or 13, characterized in that, In step 3: The priority of forward delivery, from high to low, is: emergency delivery priority, forward return and return priority, and scheduled delivery priority. Scheduled delivery priority tasks are first matched with primary transit capacity equipment. After a successful search and match, they continue to be matched with secondary terminal capacity equipment to form a transit + terminal combined delivery link. If the number of failed searches exceeds the preset search failure threshold, the search and matching will switch to secondary terminal capacity equipment to perform direct delivery. The priority of reverse pickup, from high to low, is as follows: emergency pickup priority, reverse return and return priority, and scheduled pickup priority. Scheduled pickup priority tasks are matched with secondary last-mile delivery equipment first. If the search is successful, the last-mile pickup link is marked as successfully allocated, and the search can continue to match primary transit delivery equipment. If the primary delivery equipment search is successful, a last-mile pickup + transit transportation combined link is formed. If the number of primary delivery equipment search failures exceeds the preset search failure threshold, the matched secondary last-mile delivery equipment will perform the entire pickup task.

15. The tiered hierarchical capacity scheduling and multi-objective optimization scheduling method according to claim 12 or 13, characterized in that, In step 4, the model constraints include equipment battery life, load capacity, timeliness, placement and handover time, restricted areas, total travel distance, and building access time; After the algorithm outputs the Pareto optimal solution set, it uses weighted selection based on the weight coefficients corresponding to the business scenario to select the globally optimal scheduling scheme.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 12 to 15.

Images