A Dynamic Scheduling Method, System and Medium for Logistics Vehicles

Abstract

The present invention provides a dynamic scheduling method, system and medium for logistics vehicles, which relates to the technical field of logistics management, and includes: receiving a plurality of logistics tasks to be assigned; performing hierarchical clustering to obtain M spatio-temporal coupling task groups; locally invoking the vehicle attributes of H idle logistics vehicles to perform logistics energy consumption fitting, and screening M matching logistics vehicles; solving the marginal energy consumption gain to obtain M energy consumption optimization task groups; extracting M logistics trajectories to construct M job scheduling sequences; using the M job scheduling sequences and M logistics trajectories as M physical flow dynamic scheduling strategies to drive the M matching logistics vehicles to execute a plurality of logistics tasks to be assigned. The present invention solves the technical problem in the prior art that the logistics vehicle scheduling only focuses on the task completion time and ignores the energy efficiency optimization, resulting in empty driving, overloading or inefficient path selection during the transportation process, and thus leading to excessive energy consumption.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics management, and particularly relates to a method, a system and a medium for dynamically scheduling logistics vehicles. Background Art

[0002] In large-scale logistics transportation, the energy consumption of transport vehicles not only directly affects the operating costs, but is also closely related to environmental impacts. Therefore, how to more accurately match tasks and vehicles, and on this basis, further improve efficiency by reducing energy consumption has become a technical problem to be solved urgently.

[0003] Traditional logistics vehicle scheduling methods are usually based on static rules or experience, ignoring factors such as task attributes, vehicle capabilities, and the mutual influence between tasks. Dispatchers arrange tasks and vehicles through manual planning and human decision-making, but the efficiency and flexibility of these methods are poor. Moreover, in the prior art, most intelligent scheduling methods only focus on improving scheduling efficiency or task completion time, ignoring energy consumption optimization. The inefficiency of vehicle and task allocation leads to waste of energy. For example, vehicles may be scheduled according to the time window of tasks, while ignoring the energy consumption differences of vehicles under different loads and different routes, resulting in low overall energy efficiency of the system. Summary of the Invention

[0004] This application provides a method, a system and a medium for dynamically scheduling logistics vehicles, aiming to solve the technical problem that the existing logistics vehicle scheduling only focuses on task completion time, ignores energy efficiency optimization, resulting in empty driving, overloading or inefficient path selection during transportation, and thus leading to excessive energy consumption.

[0005] In the first aspect disclosed in this application, a method for dynamically scheduling logistics vehicles is provided. The method includes: receiving a plurality of logistics tasks to be assigned, where the plurality of logistics tasks to be assigned have a plurality of task spatio-temporal attributes and a plurality of load capacity constraint attribute identifiers; hierarchically clustering the plurality of logistics tasks to be assigned according to the plurality of task spatio-temporal attributes to obtain M spatio-temporal coupled task groups; locally invoking the H vehicle attributes of H idle logistics vehicles, and performing logistics energy consumption fitting according to the H vehicle attributes and the plurality of load capacity constraint attributes to screen M matching logistics vehicles corresponding to the M spatio-temporal coupled task groups from the H idle logistics vehicles; solving the marginal energy consumption gain for the M spatio-temporal coupled task groups with the M matching logistics vehicles as constraints to obtain M energy consumption optimization task groups; extracting M logistics trajectories of the M energy consumption optimization task groups to construct M job scheduling sequences; and using the M job scheduling sequences and the M logistics trajectories as M dynamic logistics scheduling strategies to drive the M matching logistics vehicles to execute the plurality of logistics tasks to be assigned.

[0006] The second aspect disclosed in this application provides a dynamic scheduling system for logistics vehicles. The system is used for the above-mentioned dynamic scheduling method for logistics vehicles. The system includes: a logistics task receiving module, configured to receive a plurality of logistics tasks to be assigned, where the plurality of logistics tasks to be assigned have a plurality of task spatio-temporal attributes and a plurality of load capacity constraint attribute identifiers; a hierarchical clustering module, configured to perform hierarchical clustering on the plurality of logistics tasks to be assigned according to the plurality of task spatio-temporal attributes, to obtain M spatio-temporal coupled task groups; a logistics energy consumption fitting module, configured to locally call the H vehicle attributes of H idle logistics vehicles, and perform logistics energy consumption fitting according to the H vehicle attributes and the plurality of load capacity constraint attributes, so as to screen M matching logistics vehicles corresponding to the M spatio-temporal coupled task groups from the H idle logistics vehicles; an energy consumption gain solving module, configured to perform marginal energy consumption gain solving on the M spatio-temporal coupled task groups with the M matching logistics vehicles as constraints, to obtain M energy consumption optimization task groups; a scheduling sequence construction module, configured to extract the M logistics trajectories of the M energy consumption optimization task groups to construct M job scheduling sequences; a logistics task execution module, configured to use the M job scheduling sequences and the M logistics trajectories as M dynamic scheduling strategies for logistics, and drive the M matching logistics vehicles to execute the plurality of logistics tasks to be assigned.

[0007] The third aspect disclosed in this application provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the dynamic scheduling method for logistics vehicles in the first aspect are implemented.

[0008] One or more technical solutions provided in this application have at least the following beneficial effects:

[0009] Receive multiple logistics tasks to be assigned, along with their spatio-temporal attributes and load and capacity constraint attribute identifiers. The spatio-temporal attributes ensure a reasonable distribution of tasks in terms of time and space, while the load and capacity constraint attributes take into account the requirements of each task for vehicle load and volume, avoiding waste of resources and unreasonable task allocation, and providing basic data for subsequent task scheduling and optimization; perform hierarchical clustering based on the spatio-temporal attributes of the tasks to obtain M spatio-temporal coupled task groups. This grouping method can ensure the spatio-temporal coordination of tasks, making the tasks within the task groups highly similar, reducing conflicts between tasks, and optimizing the scheduling efficiency. In this way, resource utilization can be improved, idle waiting time and scheduling conflicts can be reduced; call idle logistics vehicles and their attributes, and perform logistics energy consumption fitting to screen and match logistics vehicles, achieving precise matching of tasks and vehicles, ensuring that each task group is assigned the most suitable vehicle. Through energy consumption fitting, not only the adaptability of the vehicle is considered, but also it is ensured that the vehicle can execute tasks with the lowest energy consumption, thereby improving the energy efficiency of the entire logistics system; on the basis of the matching of tasks and vehicles, solve the marginal energy consumption gain to optimize the scheduling plan, obtaining M energy consumption optimization task groups, realizing further optimization of the overall energy consumption configuration by adjusting the matching relationship between tasks and vehicles, ensuring that the execution energy efficiency of each task group is optimal. By calculating the marginal energy consumption gain, the most energy-efficient task allocation plan can be identified, improving the energy use efficiency of logistics transportation; after obtaining the optimized energy consumption optimization task groups, construct an operation scheduling sequence by extracting the logistics trajectories of each task group, ensuring the optimal execution order and path of each task, reducing empty driving and ineffective waiting during task execution, further improving vehicle utilization, and reducing transportation time and energy consumption; by combining the operation scheduling sequence and the logistics trajectory as a dynamic scheduling strategy to drive the matching logistics vehicles to execute tasks, the scheduling plan can be implemented in actual operation in real time, ensuring the dynamic execution of the scheduling plan and guaranteeing the smooth completion of logistics tasks.

[0010] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically gives the specific implementation manners of this application. Brief Description of the Drawings

[0011] Figure 1 It is a schematic flowchart of a method for dynamic scheduling of logistics vehicles provided by an embodiment of this application.

[0012] Figure 2 It is a schematic structural diagram of a system for dynamic scheduling of logistics vehicles provided by an embodiment of this application.

[0013] Description of the drawing reference numerals: Logistics task receiving module 10, hierarchical clustering module 20, logistics energy consumption fitting module 30, energy consumption gain solving module 40, scheduling sequence construction module 50, logistics task execution module 60. Detailed implementation manners

[0014] By providing a dynamic scheduling method, system and medium for logistics vehicles, the embodiment of the present application solves the technical problem in the prior art that the logistics vehicle scheduling only focuses on the task completion time and ignores the energy efficiency optimization, resulting in empty driving, overloading or inefficient path selection during the transportation process, and thus leading to excessive energy consumption.

[0015] After introducing the basic principle of the present application, the various non-limiting implementation manners of the present application will be specifically introduced below with reference to the drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0016] Embodiment 1, as Figure 1 shown, the embodiment of the present application provides a dynamic scheduling method for logistics vehicles, and the method includes:

[0017] Receiving a plurality of to-be-allocated logistics tasks, where the plurality of to-be-allocated logistics tasks have a plurality of task spatio-temporal attributes and a plurality of load capacity constraint attribute identifiers.

[0018] Receiving a plurality of to-be-allocated logistics tasks, each task having a task spatio-temporal attribute and a load capacity constraint attribute identifier. Among them, the task spatio-temporal attribute is used to describe the spatial and time characteristics of each task, and these characteristics determine the spatio-temporal overlap and mutual influence between tasks, and specifically include information such as the time window, start and end positions, and transportation conditions of the task. For example, the start and end points, start time and end time of the logistics task, etc.; the load capacity constraint attribute involves the capacity requirements of the transportation resources or containers required by the task, including the weight, volume, special transportation requirements, etc. of each task, and these all have an important impact on the subsequent vehicle matching.

[0019] According to the plurality of task spatio-temporal attributes, hierarchically clustering the plurality of to-be-allocated logistics tasks to obtain M spatio-temporal coupling task groups.

[0020] Hierarchical clustering is performed on multiple logistics tasks to be assigned according to their spatio-temporal attributes. Hierarchical clustering is a method of hierarchically dividing a dataset into multiple clusters, aiming to divide tasks into several groups, where tasks within each group have similar spatio-temporal characteristics. Specifically, the spatio-temporal overlap degree between tasks is calculated based on their spatio-temporal attributes, which measures the degree of coincidence of the spatio-temporal positions of two tasks. For example, if the time windows (such as start time and end time) and spatial attributes (starting location and destination location) of two tasks are very close, then their overlap degree is relatively high, and vice versa. After calculating the spatio-temporal overlap degree for each pair of tasks, the hierarchical clustering algorithm is used to group the tasks according to their similarity. The goal of clustering is to place tasks with high spatio-temporal similarity in the same group, thereby improving the efficiency of subsequent scheduling optimization. Through clustering, M spatio-temporally coupled task groups are finally obtained, and tasks within each group have similar spatio-temporal characteristics, facilitating subsequent vehicle matching, energy consumption optimization, and scheduling decisions.

[0021] The vehicle attributes of H idle logistics vehicles are locally called, and logistics energy consumption fitting is performed based on the H vehicle attributes and multiple load capacity constraint attributes to screen M matching logistics vehicles corresponding to the M spatio-temporally coupled task groups from the H idle logistics vehicles.

[0022] The vehicle attributes of H idle logistics vehicles are called, including load capacity, carrying volume, fuel consumption rate, driving efficiency, whether there are special functions (such as cold chain transportation), vehicle speed, etc. Logistics energy consumption fitting is performed on the H vehicle attributes and multiple load capacity constraint attributes. The goal is to estimate the energy consumed by each vehicle when performing a certain logistics task. The influencing factors of logistics energy consumption include transportation distance, transportation load, traffic conditions, vehicle performance, etc.

[0023] According to the fitting results, M matching logistics vehicles that match the M spatio-temporally coupled task groups are screened from the H idle logistics vehicles. This means that the selected vehicles can meet the time requirements of each task, that is, the vehicles can complete the tasks within the required time window; the selected vehicles can meet the load and volume requirements of each task, that is, the capacity of the vehicles is suitable for transporting the goods of the transportation tasks; at the same time, the energy consumption of the selected vehicles should be as low as possible to optimize the efficiency of the logistics process.

[0024] It should be noted that in the actual scheduling process, unforeseen situations such as vehicle failures and traffic delays may occur. Therefore, it is necessary to reserve some vehicles as spares to cope with such situations. These spare vehicles are not immediately enabled during scheduling but can be provided as alternatives when needed.

[0025] With the M matching logistics vehicles as constraints, the marginal energy consumption gain of the M spatio-temporally coupled task groups is solved to obtain M energy consumption optimization task groups.

[0026] Using the M selected matching logistics vehicles, solve the marginal energy consumption gain for each spatio-temporal coupling task group. The marginal energy consumption gain refers to the additional energy efficiency improvement that can be brought about by adjusting the matching relationship between tasks and vehicles in the scheduling. The goal of the marginal energy consumption gain is to minimize the total energy consumption, that is, by reasonably adjusting the matching of tasks and vehicles, find the optimal energy consumption configuration, so that the system consumes the least energy during the task completion process. Through the solution of the marginal energy consumption gain, M energy consumption optimization task groups are obtained. The energy consumption of these task groups has been optimized and adjusted and reached the predetermined energy consumption optimization goal. When each task group is executed, the vehicle scheduling and task arrangement can be completed with the least energy consumption, improving the overall scheduling efficiency and economy.

[0027] Extract the M logistics trajectories of the M energy consumption optimization task groups to construct M job scheduling sequences.

[0028] According to the task content of each energy consumption optimization task group, extract the corresponding logistics trajectory. The logistics trajectory refers to the specific path of the vehicle from the starting point to the ending point to execute the task, including the starting and ending locations, path selection, and the start and end times of task execution. The planning of the logistics trajectory needs to comprehensively consider the spatio-temporal attributes of the task and the driving ability of the vehicle.

[0029] Based on the extracted logistics trajectories, arrange the tasks into M job scheduling sequences according to the time sequence and vehicle capabilities. The job scheduling sequence refers to determining the task execution order of each vehicle according to factors such as logistics trajectories, vehicle attributes, and task priorities. Although each logistics vehicle has a predetermined task and path, its execution order still needs to be optimized to ensure the efficiency of the entire scheduling. Each scheduling sequence represents the specific task execution process of a vehicle, including the task execution order, the start and end times of the task, and the execution path of the vehicle.

[0030] Use the M job scheduling sequences and M logistics trajectories as M physical dynamic scheduling strategies to drive the M matching logistics vehicles to execute the multiple to-be-allocated logistics tasks.

[0031] Combine the M job scheduling sequences with the M logistics trajectories to generate M physical dynamic scheduling strategies. Finally, the M matching logistics vehicles execute their respective tasks according to these scheduling strategies. Specifically, the system will issue corresponding scheduling instructions to each vehicle, and the vehicle will execute the tasks according to the specified job scheduling sequence and logistics trajectory to ensure the efficient and low-energy consumption completion of all tasks.

[0032] Furthermore, locally call the H vehicle attributes of the H idle logistics vehicles, and perform logistics energy consumption fitting based on the H vehicle attributes and multiple load capacity constraint attributes to screen M matching logistics vehicles corresponding to the M spatio-temporal coupling task groups from the H idle logistics vehicles. The method includes:

[0033] Perform task spatio-temporal attribute decomposition on the first spatio-temporal coupling task group to extract R logistics starting and ending points that make up the tasks; perform shortest logistics route fitting based on the R logistics starting and ending points to obtain the first logistics trajectory and the first operation scheduling sequence; according to the R tasks that make up the tasks, call R load capacities from the multiple load capacity constraint attributes; according to the H vehicle attributes, the first operation scheduling sequence, the first logistics trajectory, and the R load capacities, perform global logistics energy consumption simulation on the H idle logistics vehicles to obtain the first group of logistics energy consumption information, where the first group of logistics energy consumption information includes H first global logistics energy consumptions; and so on, through logistics energy consumption fitting, obtain the M groups of logistics energy consumption information of the M spatio-temporal coupling task groups; with the goal of minimizing energy consumption, perform allocation conflict resolution on the H idle logistics vehicles according to the M groups of logistics energy consumption information, and output the M matching logistics vehicles.

[0034] Multiple logistics tasks have been hierarchically clustered by spatio-temporal attributes to form M spatio-temporal coupling task groups. In this step, further analysis is performed on the first spatio-temporal coupling task group among them. Specifically, spatio-temporal attribute decomposition decomposes each task of the first spatio-temporal coupling task group into the starting and ending points that make up the task, and these starting and ending points include: the starting point of the task, such as the origin of the goods; the ending point of the task, such as the destination of the goods; and the transfer points between each task, such as intermediate stops or transfer points. When performing spatio-temporal attribute decomposition, based on the spatial and temporal requirements of the task, R logistics starting and ending points are extracted, representing the starting and ending positions of each subtask in the first spatio-temporal coupling task group, and these positions will be used as the basis for subsequent path planning and scheduling.

[0035] Based on the extracted R logistics starting and ending points, perform shortest logistics route fitting. The goal is to calculate the shortest logistics route according to the geographical location and traffic information between the starting and ending points. During the fitting process, there may be multiple paths that can connect between each starting and ending point, and the shortest path is selected. According to the fitted shortest logistics route, generate the first logistics trajectory, which represents the specific route of the vehicle from one starting and ending point to another when executing this spatio-temporal coupling task group. At the same time, generate the first operation scheduling sequence based on the logistics trajectory. The operation scheduling sequence determines the order of each vehicle when executing the task, as well as the time and location of each task execution.

[0036] When each task is executed, there are certain capacity constraints, such as the weight, volume of goods or other special requirements. These constraints determine whether the vehicle can carry the task. According to the specific information of the R constituent tasks, match with the corresponding capacity constraint attributes, and extract the corresponding R capacity masses, which include the transportation resources required for each task (such as load capacity, volume requirements, etc.) and the adaptation ability of the vehicle. Through these capacity masses, it can be determined which vehicles can meet the transportation requirements of the corresponding tasks. If a vehicle has insufficient load capacity or cannot meet special transportation requirements, then it will be excluded from the optional vehicles.

[0037] The global logistics energy consumption simulation is a comprehensive calculation process based on multiple factors. In this step, combine H vehicle attributes, the first operation scheduling sequence, the first logistics trajectory and R capacity masses to simulate the energy consumption of each vehicle when performing tasks. Specifically, combine the above factors, and by simulating each vehicle, calculate the global logistics energy consumption of each vehicle when performing specific tasks. The global logistics energy consumption refers to the total energy consumed by the vehicle during the entire task process. This energy consumption includes the energy consumption in various aspects such as vehicle driving, loading, road conditions, and parking waiting. Finally, obtain the first set of logistics energy consumption information, which includes the energy consumption information of each idle vehicle when performing this task, that is, H first global logistics energy consumptions, and each value represents the energy consumption of each vehicle to complete the task.

[0038] Repeat the above process to perform energy consumption simulation on each task group of the M spatio-temporal coupling task groups. Finally, obtain M sets of logistics energy consumption information. These energy consumption information reflect the energy consumption characteristics of each task group and are used to further optimize the scheduling process. For example, if the energy consumption of a certain task group is relatively high, then it is necessary to adjust vehicle allocation, task sequence or path to reduce energy consumption.

[0039] With the goal of minimizing energy consumption, try to reduce the overall energy consumption by optimizing the matching between tasks and vehicles. Through the analysis of the M sets of logistics energy consumption information, ensure that the allocation of tasks among vehicles can minimize energy consumption. Among them, allocation conflict means that it is possible that a vehicle has the lowest energy consumption in two or more spatio-temporal coupling task groups. To avoid a vehicle being repeatedly allocated, it is necessary to resolve allocation conflicts. The ways to resolve allocation conflicts include: adjusting task allocation, removing some tasks from the current vehicle and reallocating them to other vehicles with lower energy consumption; priority sorting, according to the timeliness, importance or other criteria of tasks, preferentially allocate vehicles for some tasks; standby vehicle scheduling, rescheduling some tasks to standby vehicles, so as to avoid repeated allocation of vehicles.

[0040] After completing the allocation conflict resolution, M matching logistics vehicles are output. These vehicles are assigned to M spatio-temporal coupling task groups, and the energy consumption during task execution is optimized. Through conflict resolution and optimization, it is ensured that each task group can obtain the most suitable vehicle, while avoiding duplicate allocation, ensuring scheduling efficiency and energy efficiency.

[0041] Furthermore, with the M matching logistics vehicles as constraints, the marginal energy consumption gain of the M spatio-temporal coupling task groups is solved to obtain M energy consumption optimization task groups. The method includes:

[0042] According to the M matching logistics vehicles, M logistics energy consumption baselines are extracted from the M sets of logistics energy consumption information; according to the spatio-temporal attributes of the multiple tasks, the spatio-temporal overlap degree of the M spatio-temporal coupling task groups is calculated to screen out M initial boundary tasks; according to the multiple load capacity constraint attributes, the task similarity of the M initial boundary tasks is calculated, and the M initial boundary tasks are combined according to the calculation results to obtain a candidate task adjustment pair; according to the a candidate task adjustment pair, inter-group task migration is performed on the M spatio-temporal coupling task groups, and the marginal energy consumption gain is solved according to the M matching logistics vehicles to obtain M migrated task groups; and so on. After calculating the spatio-temporal overlap degree of the M migrated task groups according to the spatio-temporal attributes of the multiple tasks to screen out M updated boundary tasks, inter-group task migration and marginal energy consumption gain solution are performed to update the M migrated task groups until the M energy consumption optimization task groups with an energy consumption gain less than a preset threshold are obtained through iterative solution of the marginal energy consumption gain.

[0043] M logistics energy consumption baselines are extracted from the logistics energy consumption information of each matching vehicle. The logistics energy consumption baseline refers to the standard energy consumption value of each vehicle when executing a set of tasks. It represents the energy consumption reference value when the vehicle executes tasks. Usually, it is the basic energy consumption of the vehicle when executing tasks without any optimization adjustments.

[0044] According to the spatio-temporal attributes of each task (including the time window, geographical location, transportation route, etc.) of the task, the spatio-temporal overlap degree is calculated. The spatio-temporal overlap degree refers to the coincidence degree between the spatio-temporal characteristics of two tasks, including the time overlap degree and the space overlap degree. Based on the calculated spatio-temporal overlap degree, M initial boundary tasks are screened out. These tasks refer to the boundary tasks with the lowest spatio-temporal overlap degree with other tasks in the same spatio-temporal coupling task group. This means that these tasks are relatively independent of other tasks in the group in terms of time and space and have high flexibility for adjustment or migration. Through such a selection, it can be ensured that subsequent task migration and optimization will not damage the spatio-temporal coordination within the task group, and at the same time provide space for further adjustment of the tasks.

[0045] According to the load capacity constraint attributes of each task, such as the weight and volume of the task, calculate the task similarity for M initial boundary tasks. Task similarity measures the similarity degree of two tasks in terms of load capacity requirements. The calculation factors include load requirements, volume requirements, and special transportation requirements. For example, if the load requirements of two tasks are similar, their similarity is relatively high; if the transportation space required by two tasks is similar, their similarity is also relatively high.

[0046] Based on the calculated task similarity, combine those initial boundary tasks with relatively high similarity and generate candidate task adjustment pairs. Each adjustment pair consists of two initial boundary tasks with relatively high similarity, and they have high similarity in terms of load capacity requirements and spatio-temporal coordination. represents the number of different adjustment pairs existing between tasks. This number is derived through combinatorial mathematics without considering duplicate combinations. These task adjustment pairs represent task pairs that can be migrated between task groups.

[0047] According to candidate task adjustment pairs, perform inter-group task migration. Inter-group task migration refers to transferring tasks in a spatio-temporal coupled task group to another spatio-temporal coupled task group to optimize task scheduling and reduce energy consumption. The marginal energy consumption gain refers to the optimization effect on energy consumption after adjusting task allocation. When performing task migration, recalculate the energy consumption of the task group after migration. Specifically, after task migration, the vehicle needs to execute new tasks and routes, so it is necessary to calculate the new energy consumption according to the energy change brought by the migrated tasks. According to the new task allocation, calculate the energy consumption of the vehicle when executing these new tasks and compare it with the energy consumption before migration to obtain the marginal energy consumption gain. The goal of the marginal energy consumption gain is to minimize energy consumption, that is, by migrating tasks and adjusting vehicle scheduling, reduce the overall energy consumption. After task migration and marginal energy consumption gain calculation, M migrated task groups are obtained. These task groups represent the optimized task allocation, and their energy consumption has been improved.

[0048] After obtaining M migrated task groups, continue to calculate the spatio-temporal overlap degree and screen out M updated boundary tasks based on the calculation results. These updated boundary tasks refer to the tasks with the lowest spatio-temporal overlap degree with other tasks within the task group, and they can be flexibly migrated in subsequent iterations to further optimize scheduling. After screening out the updated boundary tasks, perform inter-group task migration again. The migrated task group will perform further marginal energy consumption gain solution according to the energy consumption of the vehicle. Continue to iterate the energy consumption of the task group by migrating tasks and recalculating the energy consumption. The goal of the iterative process is to gradually reduce energy consumption through multiple rounds of adjustment and finally converge to a stable energy consumption level, so that the optimized task group can be completed with as low energy consumption as possible.

[0049] After each task migration and energy consumption solution, check whether the marginal energy consumption gain is less than a preset threshold. The preset threshold is set according to the actual situation and specific requirements. If it is less than the preset threshold, the iterative process will stop, and the final M energy consumption optimization task groups will be obtained. These final task groups can provide the optimal energy efficiency and task allocation scheme after multiple rounds of adjustment, ensuring that the vehicle can complete tasks efficiently and reduce energy waste.

[0050] Furthermore, with the M matching logistics vehicles as constraints, solve the marginal energy consumption gain for the M spatio-temporal coupling task groups to obtain M energy consumption optimization task groups. The method includes:

[0051] According to the first initial boundary task and the second initial boundary task that form the first candidate task adjustment pair, locate the second spatio-temporal coupling task group and the third spatio-temporal coupling task group in the M spatio-temporal coupling task groups. Among them, the second spatio-temporal coupling task group and the third spatio-temporal coupling task group are processed by the first matching logistics vehicle and the second matching logistics vehicle respectively; perform task exchange migration of the first initial boundary task and the second initial boundary task between the second spatio-temporal coupling task group and the third spatio-temporal coupling task group to obtain the second migration task group and the third migration task group; respectively perform logistics energy consumption fitting on the second migration task group and the third migration task group according to the first matching logistics vehicle and the second matching logistics vehicle to obtain the first migration logistics energy consumption and the second migration logistics energy consumption; extract the first logistics energy consumption baseline and the second logistics energy consumption baseline from the M logistics energy consumption baselines according to the second spatio-temporal coupling task group and the third spatio-temporal coupling task group; if the marginal energy consumption gain obtained by adding the first migration logistics energy consumption and the second migration logistics energy consumption is better than the marginal energy consumption gain obtained by adding the first logistics energy consumption baseline and the second logistics energy consumption baseline, then retain the second migration task group and the third migration task group; and so on, perform inter-group task migration according to the candidate task adjustments, and solve the marginal energy consumption gain according to the M matching logistics vehicles to obtain the M migration task groups.

[0052] In the obtained Among the candidate task adjustment pairs, randomly select an analysis object as the first candidate task adjustment pair. This first candidate task adjustment pair includes a first initial boundary task and a second initial boundary task. The first initial boundary task and the second initial boundary task come from any two of the M spatio-temporal coupling task groups respectively. Perform task group localization to obtain a second spatio-temporal coupling task group and a third spatio-temporal coupling task group, which are the groups to which the first initial boundary task and the second initial boundary task belong respectively. The second spatio-temporal coupling task group and the third spatio-temporal coupling task group are processed using different matching logistics vehicles. Specifically, the second spatio-temporal coupling task group uses a first matching logistics vehicle, and the third spatio-temporal coupling task group uses a second matching logistics vehicle.

[0053] Perform task migration and exchange between the second spatio-temporal coupling task group and the third spatio-temporal coupling task group, that is, migrate the first initial boundary task in the second spatio-temporal coupling task group to the third spatio-temporal coupling task group, and at the same time migrate the second initial boundary task in the third spatio-temporal coupling task group to the second spatio-temporal coupling task group. After the migration is completed, a new task combination is formed: the second spatio-temporal coupling task group removes the first initial boundary task and adds the second initial boundary task, serving as the second migration task group; the third spatio-temporal coupling task group removes the second initial boundary task and adds the first initial boundary task, serving as the third migration task group. The tasks in these task groups have been reallocated.

[0054] Perform logistics energy consumption fitting on the second migration task group and the third migration task group respectively. The goal of logistics energy consumption fitting is to estimate the energy consumption of vehicles when performing these tasks based on the attributes of each vehicle in the task group and the characteristics of task execution. Among them, the first matching logistics vehicle is responsible for executing the second migration task group, and the second matching logistics vehicle is responsible for executing the third migration task group. The influencing factors of logistics energy consumption fitting include: the load and volume requirements of the tasks. The scale of the tasks will affect the energy consumption of the vehicles, especially for tasks with heavy loads or special transportation conditions (such as cold chain); route selection, that is, the specific path of each migration task group, combined with factors such as traffic conditions and road types; vehicle attributes, that is, the performance of each matching vehicle, including fuel efficiency, power consumption rate, driving efficiency, etc.; time window. The time requirements for task execution will also affect the energy consumption, because tasks with shorter execution times or stronger timeliness may lead to increased energy consumption.

[0055] By performing energy consumption fitting on each task group, calculate the first migration logistics energy consumption and the second migration logistics energy consumption, which respectively represent the total energy consumption when the first matching logistics vehicle executes the second migration task group and the second matching logistics vehicle executes the third migration task group.

[0056] Based on the task situations of the second spatio-temporal coupling task group and the third spatio-temporal coupling task group, the corresponding first logistics energy consumption baseline and second logistics energy consumption baseline are extracted from M baselines. The first logistics energy consumption baseline represents the standard energy consumption when the first matching logistics vehicle executes the second spatio-temporal coupling task group under default conditions; the second logistics energy consumption baseline represents the standard energy consumption when the second matching logistics vehicle executes the third spatio-temporal coupling task group under default conditions. These logistics energy consumption baselines reflect the energy efficiency reference values of the vehicle when performing specific tasks.

[0057] Add the first migrated logistics energy consumption and the second migrated logistics energy consumption to obtain the total energy consumption of the migration task group. Then, add the first logistics energy consumption baseline and the second logistics energy consumption baseline to obtain the total energy consumption of the baseline task group. Compare the two total energy consumptions. If the total energy consumption of the migration task group is lower, it indicates that the task migration optimization is effective and the energy efficiency has been improved. In this case, retain the second migrated task group and the third migrated task group after migration and use them as the new scheduling plan. Conversely, if the energy consumption after migration is not better than the baseline energy consumption, withdraw the task migration and continue to search for a better task group configuration.

[0058] Repeat the above process for more inter-group task migrations of candidate tasks. These candidate task adjustments represent possible options for migration between different task groups. After each task migration, solve the marginal energy consumption gain for the newly migrated task group based on M matching logistics vehicles. The goal is to ensure the minimization of energy consumption by continuously adjusting the matching between tasks and vehicles, calculate the energy efficiency improvement of each migrated task group compared to before migration, until the optimal configuration is found. Finally, obtain M migrated task groups. These task groups represent the optimized scheduling plan, and the energy efficiency of each task group will be as low as possible to ensure the operation efficiency and sustainability of the entire logistics system.

[0059] Furthermore, according to the multiple task spatio-temporal attributes, perform hierarchical clustering on the multiple logistics tasks to be assigned to obtain M spatio-temporal coupling task groups. The method includes:

[0060] Combined enumerate the multiple logistics tasks to be assigned to obtain W pairs of logistics tasks to be assigned; refer to the W pairs of logistics tasks to be assigned and combine the multiple task spatio-temporal attributes to obtain W pairs of task spatio-temporal attributes; calculate the spatio-temporal overlap degree according to the W pairs of task spatio-temporal attributes to obtain W task spatio-temporal overlap degrees; perform hierarchical clustering on the multiple logistics tasks to be assigned according to the W task spatio-temporal overlap degrees to obtain the M spatio-temporal coupling task groups.

[0061] Perform combinatorial enumeration on multiple to-be-allocated logistics tasks. Combinatorial enumeration involves selecting pairs of tasks from all to-be-allocated tasks to form task pairs, that is, W to-be-allocated logistics tasks. Specifically, by selecting different task pairs from multiple to-be-allocated logistics tasks, possible task combinations are generated. Suppose there are N to-be-allocated tasks, then all possible pairwise combinations of these tasks are obtained, resulting in W pairs of tasks, where , which is derived through combinatorial mathematics without considering duplicate combinations.

[0062] Each to-be-allocated logistics task has corresponding task spatio-temporal attributes, including information such as the time window, start and end positions, and transportation conditions of the task. Combine the corresponding spatio-temporal attributes of W to-be-allocated logistics tasks. For example, for any task pair, which includes Task 1 and Task 2, refer to their time windows and spatial positions to determine their spatio-temporal relationship, such as whether the time windows of Task 1 and Task 2 overlap, whether their starting and ending points are close, or whether their transportation conditions are similar, etc. Through these combinations, corresponding W pairs of task spatio-temporal attributes are generated for the W pairs of tasks, and these attributes are used for subsequent spatio-temporal overlap degree calculations.

[0063] The spatio-temporal overlap degree measures the degree of overlap between two tasks in terms of time and space, including the time overlap degree and the space overlap degree. Among them, the time overlap degree is reflected in whether the time windows of the tasks overlap, and the longer the overlapping time length, the higher the overlap degree. For example, if the time windows of two tasks have an intersection, then their time overlap degree is relatively high; the space overlap degree is reflected in whether the starting and ending points of the tasks are close, or whether the transportation routes involved in the tasks overlap. The space overlap degree is usually calculated based on the geographical locations of the tasks. The smaller the spatial distance between tasks, the higher the overlap degree. Calculate the spatio-temporal overlap degree according to the spatio-temporal attributes of each pair of tasks, and finally obtain W task spatio-temporal overlap degrees, indicating the degree of spatio-temporal overlap of each pair of tasks.

[0064] Based on the calculated W task spatio-temporal overlap degrees, perform hierarchical clustering on multiple to-be-allocated logistics tasks. Hierarchical clustering is an algorithm for grouping based on the similarity between data, used to classify data according to hierarchical relationships. In this context, the spatio-temporal overlap degree of tasks reflects the similarity between tasks. By using the spatio-temporal overlap degree to determine which tasks should be assigned to the same group and which tasks should be assigned to different groups. If the spatio-temporal overlap degree of two tasks is relatively high, then they are assigned to the same group; if the spatio-temporal overlap degree is relatively low, then they are assigned to different groups. During the process of hierarchical clustering, tasks with relatively high spatio-temporal overlap degrees are aggregated together to form spatio-temporal coupled task groups one by one. Finally, M spatio-temporal coupled task groups are obtained. The tasks within these task groups have relatively high spatio-temporal similarity and can share resources or be carried out simultaneously during execution.

[0065] Furthermore, calculate the spatio-temporal overlap degree of the task spatio-temporal attributes according to W, obtaining W task spatio-temporal overlap degrees. The method includes:

[0066] Decompose the W pairs of task spatio-temporal attributes to obtain W pairs of logistics start and end points and W pairs of logistics time windows; locally call the task transportation mean value, and calculate the maximum reusable distance according to the task transportation mean value and the W pairs of logistics start and end points, and output W logistics distance overlap degrees; calculate the window intersection percentage of the W pairs of logistics time windows to obtain W time window overlap degrees; pre-configure the logistics weight relationship to perform weighted fusion of the W logistics distance overlap degrees and the W time window overlap degrees to obtain the W task spatio-temporal overlap degrees.

[0067] Decompose the W pairs of task spatio-temporal attributes. The spatio-temporal attribute of each task includes the logistics start and end points and the logistics time window. Among them, the logistics start and end points refer to the starting point (such as the loading location of the goods) and the ending point (such as the unloading location of the goods) of each task, and the logistics time window refers to the start time and end time of the task, representing the time range for task execution. Decompose to obtain W pairs of logistics start and end points, representing the starting and ending points of each pair of tasks, and W pairs of logistics time windows, representing the start and end times of each pair of tasks.

[0068] Locally call the task transportation mean value. The transportation mean value includes the transportation distance and transportation time, which represents the average transportation distance and time of tasks in the system. According to the W pairs of logistics start and end points and the transportation mean value, calculate the maximum reusable distance. This distance represents the maximum path part that can be shared between two tasks, and the calculation formula is as follows:

[0069] ;

[0070] Among them, characterizes the task and the task 's maximum reusable distance, characterizes the task the straight-line distance from the end point to the start point of the task , characterizes the average transportation distance of all tasks in the region, calculated based on historical data statistical values, and is used for normalization.

[0071] Exemplarily, if the distance from the end point of task A to the start point of task B is 80 KM and the regional average distance is 100 KM, then the maximum reusable distance between task A and task B .

[0072] After the above maximum reusable distance calculation, output W logistics distance overlap degrees. Each overlap degree represents the spatial overlap degree between task pairs, that is, the proportion of their path parts that can be reused.

[0073] Each task has a designated time window, which represents the start time and end time of the task. For a pair of tasks, calculate whether their time windows have an intersection. If the time windows of task i and task j have an intersection, then there is an overlapping period. The larger the intersection, the higher the degree of time overlap between the tasks. For task i and task j, calculate the intersection part of their time windows and compare it with the total length of their time windows to obtain the proportion of the intersection, which is used as the degree of overlap of their time windows. By analogy, by calculating the intersections of the time windows of all task pairs, W degrees of overlap of time windows are obtained. Each overlap value represents the degree of time overlap between a pair of tasks, and the larger it is, the stronger the time overlap between the tasks.

[0074] Pre-configure the logistics weight relationship, which consists of spatial weight and temporal weight. The value of each weight is between [0,1], and the sum of the two is 1. This relationship determines the importance of spatial overlap and temporal overlap in calculation. The allocation of weights can be determined based on business requirements. For example, in some applications, more emphasis is placed on temporal overlap, so the weight of temporal overlap is higher.

[0075] According to the spatial weight and temporal weight, perform weighted fusion on the W degrees of logistics distance overlap and the W degrees of time window overlap. The purpose of weighted fusion is to comprehensively consider the spatio-temporal overlap of tasks and obtain a more comprehensive degree of spatio-temporal overlap of tasks, so as to better reflect the overall similarity between tasks. Through the above weighted fusion, finally W degrees of spatio-temporal overlap of tasks are obtained, and these spatio-temporal overlaps reflect the overall spatio-temporal overlap degree of each pair of tasks.

[0076] Furthermore, based on the W degrees of spatio-temporal overlap of tasks, perform hierarchical clustering on the multiple logistics tasks to be allocated to obtain the M spatio-temporal coupling task groups. The method includes:

[0077] Construct a symmetric matrix based on the W degrees of spatio-temporal overlap of tasks and the W pairs of logistics tasks to be allocated to obtain a task similarity matrix; locally call the logistics load constraints, where the logistics load constraints include the lower limit of logistics quality, the lower limit of logistics volume, and the lower limit of vehicle redundancy; based on the logistics load constraints and multiple load capacity constraint attributes, perform multiple rounds of descending cluster merging on the task similarity matrix and output the M spatio-temporal coupling task groups.

[0078] The spatio-temporal overlap degree reflects the spatio-temporal similarity between tasks. A higher overlap degree indicates a stronger spatio-temporal similarity between tasks during execution, and tasks are more suitable to be assigned to the same task group. Use the spatio-temporal overlap degrees of W tasks to construct a symmetric matrix. Each element of the symmetric matrix represents the overlap degree between two tasks. Specifically, the rows and columns of the symmetric matrix represent different tasks, and each element of the matrix represents the spatio-temporal overlap degree between the corresponding task i and task j. This matrix can be filled with the spatio-temporal overlap degrees between tasks. For example, if the time window overlap degree between task i and task j is high and their starting and ending points are close, the value at the corresponding position in the matrix will be large, indicating a high similarity between them. In this way, the task similarity matrix is finally obtained, which represents the spatio-temporal similarity between all pairs of tasks to be assigned.

[0079] Invoke the logistics load constraint to ensure that task assignment can guarantee the suitability of the task group, avoid micro tasks forming a group alone, and improve vehicle utilization. The logistics load constraint includes the lower limit of logistics quality, the lower limit of logistics volume, and the lower limit of vehicle redundancy. Among them, the lower limit of logistics quality is the minimum load required for each task group. If the total goods of a certain task group are light, then it is not suitable to form a group alone and needs to be merged with other tasks to improve the load efficiency of the vehicle; the lower limit of logistics volume is the minimum transportation volume required for each task group. If the total volume of goods of the task group is small, then multiple small-volume tasks need to be combined to improve vehicle utilization. These two conditions ensure that tasks will not form a group alone during clustering, but consider the actual load and volume requirements of tasks, avoid assigning too small tasks alone, and thus improve vehicle utilization. The lower limit of vehicle redundancy is used to limit the redundant number of vehicles, which represents the difference between the total number of vehicles and the number of idle vehicles. Redundant vehicles are used to support dynamic task insertion and fault tolerance.

[0080] This logistics load constraint helps to avoid the situation of micro tasks forming a group alone, thereby improving vehicle utilization. By merging suitable tasks into the same group, resource utilization efficiency can be improved, and empty running or underutilization can be reduced.

[0081] Perform multi-round descending cluster merging. Specifically, the elements of the task similarity matrix represent the similarity between tasks. First, sort according to the similarity values and preferentially merge those tasks with high similarity. During the merging process, gradually merge the task clusters until the logistics load constraints are reached, including the lower limit of logistics load, the lower limit of vehicle redundancy, etc. Each round of merging is based on the similarity between tasks and at the same time considers the load constraints. The merging process will ensure that each merged cluster meets the constraint conditions in terms of spatio-temporal, load, volume, and vehicle redundancy. After multi-round descending cluster merging, finally M spatio-temporally coupled task groups are obtained. The tasks within these task groups have high spatio-temporal similarity and meet the constraint requirements in terms of load, volume, and the number of vehicles, avoiding empty running or other wasteful situations.

[0082] Furthermore, based on the logistics load constraint and multiple load capacity constraint attributes, multiple rounds of descending cluster merging are performed on the task similarity matrix to output the M spatio-temporal coupling task groups. The method includes:

[0083] Taking the lower limit of logistics quality and the lower limit of logistics volume as cluster division conditions, combining the multiple load capacity constraint attributes, randomly dividing task clusters in the task similarity matrix to obtain K random clusters; after calculating the average inter-cluster task similarity after combinatorial enumeration of the K random clusters, taking the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combining the similarity ranking to aggregate the K random clusters into Q initial aggregated clusters; after calculating the average inter-cluster task similarity after combinatorial enumeration of the Q initial aggregated clusters, taking the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combining the similarity ranking to aggregate the Q initial aggregated clusters into F second aggregated clusters; and so on, performing multiple rounds of descending cluster merging based on similarity until the number of clusters is reduced to meet the vehicle redundancy lower limit, and extracting the M spatio-temporal coupling task groups, where K > Q > F > M, and K, Q, F, and M are positive integers.

[0084] The lower limit of logistics quality and the lower limit of logistics volume ensure that tasks will not form a separate group during clustering. Instead, considering the actual load and volume requirements of the tasks, it avoids allocating too small tasks separately, thereby improving the utilization rate of vehicles. According to the lower limit of logistics quality, the lower limit of logistics volume, and the load capacity constraint attributes of the tasks, the tasks are divided into K random clusters. Each cluster contains some tasks that belong to the same category under load, volume, and other constraint conditions. However, these clusters are not the final optimized results because there may still be differences in the similarity between tasks.

[0085] The random division here means that tasks are not allocated in strict spatio-temporal similarity in advance, but are randomly grouped according to the basic load requirements of the tasks. The purpose of this random division step is to roughly allocate tasks in the initial stage without immediately enforcing the optimal task combination. The random division provides a basis for subsequent aggregation and optimization.

[0086] Perform combinatorial enumeration on the K random clusters. Combinatorial enumeration refers to performing combinatorial analysis between random clusters and evaluating the task similarity between each pair of clusters. The task similarity is calculated based on the spatio-temporal attributes of the tasks, including time window overlap, spatial distance, etc. If the tasks in two clusters highly overlap spatio-temporally, the task similarity is high.

[0087] After calculating the similarity between tasks in different clusters, the lower limit of logistics quality and the lower limit of logistics volume are used as the cluster aggregation conditions. That is, it does not simply rely on task similarity to merge clusters, but also considers the load requirements of each cluster simultaneously to ensure that the aggregated clusters can meet the load and volume requirements. Under this cluster aggregation condition, according to the similarity ranking of task clusters, clusters with higher similarity are preferentially merged together, which helps to reduce the differences between different clusters and improve the coordination within the task group. After the similarity ranking and the screening of aggregation conditions, K random clusters are merged into Q initial aggregated clusters. The tasks within each aggregated cluster are relatively similar in terms of time and space and meet the requirements in terms of load constraints such as load and volume.

[0088] Further optimization is carried out on the Q initial aggregated clusters obtained. By means of combinatorial enumeration, the task similarity between different initial aggregated clusters is calculated. The calculation of task similarity is still based on the spatio-temporal attributes of tasks, including the overlap of time windows, spatial distance, etc.

[0089] After calculating the average task similarity between clusters, the lower limit of logistics quality and the lower limit of logistics volume are combined as the aggregation conditions to ensure that the merged task clusters can meet the load and volume requirements of each task. Under this cluster aggregation condition, according to the task similarity ranking, task clusters with higher similarity are preferentially merged. The merged clusters need to meet the load constraints to ensure that the tasks within each cluster can be executed efficiently and without conflicts. Through the above similarity calculation and load constraint conditions, the Q initial aggregated clusters are merged into F second aggregated clusters. The tasks in the second aggregated clusters still maintain a high spatio-temporal similarity and meet the constraints such as load and volume.

[0090] After obtaining the F second aggregated clusters, multiple rounds of descending-order cluster merging are continued. This process is iterative. In each round of merging, clusters with higher similarity are merged together. The goal is to reduce the number of clusters while improving the task coordination and energy efficiency of each cluster. During the process of cluster merging, it is continuously checked whether each merged cluster meets the lower limit of vehicle redundancy. The lower limit of vehicle redundancy refers to the number of redundant vehicles reserved in the scheduling, which ensures that even when some vehicles break down or unforeseen situations occur, there are enough spare vehicles to handle these emergencies and continue to execute tasks.

[0091] When the lower limit of vehicle redundancy is met, M spatio-temporal coupled task groups are extracted. These task groups represent the optimal task allocation. The tasks within each task group have a high spatio-temporal similarity and meet various constraints such as load, volume, and redundancy. Throughout the process, the number of clusters is gradually reduced. Initially, there are K clusters, then it is reduced to Q clusters, then to F clusters, and finally M spatio-temporal coupled task groups are obtained. This decreasing process reflects the gradual refinement of task allocation and optimization.

[0092] In summary, the dynamic scheduling method for logistics vehicles provided by the embodiments of the present application has the following technical effects:

[0093] Receive multiple logistics tasks to be allocated and their spatio-temporal attributes and load capacity constraint attribute identifiers. The spatio-temporal attributes ensure the reasonable distribution of tasks in terms of time and space, while the load capacity constraint attributes consider the requirements of each task for vehicle load and volume, avoiding waste of resources and unreasonable task allocation, and providing basic data for subsequent task scheduling and optimization; perform hierarchical clustering based on the task spatio-temporal attributes to obtain M spatio-temporal coupled task groups. This grouping method can ensure the spatio-temporal coordination of tasks, making the tasks within the task group highly similar, reducing conflicts between tasks, and optimizing the scheduling efficiency. In this way, resource utilization can be improved, idle waiting time and scheduling conflicts can be reduced; call idle logistics vehicles and their attributes, and perform logistics energy consumption fitting to screen and match logistics vehicles, achieving precise matching of tasks and vehicles, ensuring that each task group is assigned the most suitable vehicle. Through energy consumption fitting, not only the adaptability of the vehicle is considered, but also the vehicle can execute tasks with the lowest energy consumption, thereby improving the energy efficiency of the entire logistics system; on the basis of the matching of tasks and vehicles, solve the marginal energy consumption gain to optimize the scheduling plan, and obtain M energy consumption optimization task groups, realizing the further optimization of the overall energy consumption configuration by adjusting the matching relationship between tasks and vehicles, ensuring the optimal execution energy efficiency of each task group. By calculating the marginal energy consumption gain, the most energy-efficient task allocation plan can be identified, improving the energy use efficiency of logistics transportation; after obtaining the optimized energy consumption optimization task groups, construct an operation scheduling sequence by extracting the logistics trajectories of each task group, ensuring the optimal execution order and path of each task, reducing empty driving and ineffective waiting during task execution, further improving vehicle utilization, and reducing transportation time and energy consumption; by combining the operation scheduling sequence and the logistics trajectory as a dynamic scheduling strategy to drive the matching logistics vehicles to execute tasks, the scheduling plan can be implemented in actual operation in real time, ensuring the dynamic execution of the scheduling plan and guaranteeing the smooth completion of logistics tasks.

[0094] Embodiment 2, based on the same inventive concept as the dynamic scheduling method for logistics vehicles in the foregoing embodiment, as Figure 2 shown, the embodiments of the present application provide a dynamic scheduling system for logistics vehicles, and the system includes:

[0095] A logistics task receiving module 10, configured to receive multiple logistics tasks to be allocated, and the multiple logistics tasks to be allocated have multiple task spatio-temporal attributes and multiple load capacity constraint attribute identifiers.

[0096] A hierarchical clustering module 20, configured to perform hierarchical clustering on the multiple logistics tasks to be allocated according to the multiple task spatio-temporal attributes to obtain M spatio-temporal coupled task groups.

[0097] The logistics energy consumption fitting module 30 is used to locally call the H vehicle attributes of H idle logistics vehicles, and perform logistics energy consumption fitting based on the H vehicle attributes and multiple load capacity constraint attributes, so as to screen M matching logistics vehicles corresponding to the M spatio-temporal coupling task groups from the H idle logistics vehicles.

[0098] The energy consumption gain solving module 40 is used to perform marginal energy consumption gain solving on the M spatio-temporal coupling task groups with the M matching logistics vehicles as constraints, and obtain M energy consumption optimization task groups.

[0099] The scheduling sequence construction module 50 is used to extract the M logistics trajectories of the M energy consumption optimization task groups and construct M job scheduling sequences.

[0100] The logistics task execution module 60 is used to use the M job scheduling sequences and M logistics trajectories as M physical dynamic scheduling strategies to drive the M matching logistics vehicles to execute the multiple to-be-allocated logistics tasks.

[0101] Furthermore, the logistics energy consumption fitting module 30 is used to perform the following operation steps:

[0102] Perform task spatio-temporal attribute decomposition on the first spatio-temporal coupling task group to extract R logistics start and end points of R constituent tasks; perform the shortest logistics route fitting based on the R logistics start and end points to obtain the first logistics trajectory and the first job scheduling sequence; call R load capacities from the multiple load capacity constraint attributes according to the R constituent tasks; perform global logistics energy consumption simulation on the H idle logistics vehicles according to the H vehicle attributes, the first job scheduling sequence, the first logistics trajectory, and the R load capacities to obtain the first set of logistics energy consumption information, where the first set of logistics energy consumption information includes H first global logistics energy consumptions; and so on, through logistics energy consumption fitting, obtain the M sets of logistics energy consumption information of the M spatio-temporal coupling task groups; with the goal of minimizing energy consumption, perform allocation conflict resolution on the H idle logistics vehicles according to the M sets of logistics energy consumption information, and output the M matching logistics vehicles.

[0103] Furthermore, the energy consumption gain solving module 40 is used to perform the following operation steps:

[0104] According to the M matching logistics vehicles, extract M logistics energy consumption baselines from the M sets of logistics energy consumption information; calculate the spatio-temporal overlap degree of the M spatio-temporal coupling task groups according to the multiple task spatio-temporal attributes to screen out M initial boundary tasks; calculate the task similarity of the M initial boundary tasks according to the multiple load capacity constraint attributes, and combine the M initial boundary tasks according to the calculation results to obtain Candidate task adjustment pairs; according to the For each candidate task adjustment, perform inter-group task migration on the M spatio-temporal coupled task groups, and solve for the marginal energy consumption gain based on the M matching logistics vehicles to obtain M migrated task groups; and so on. After calculating the spatio-temporal overlap degree of the M migrated task groups according to the multiple task spatio-temporal attributes to screen out M updated boundary tasks, perform inter-group task migration and marginal energy consumption gain solution to update the M migrated task groups until the M energy consumption optimization task groups with an energy consumption gain less than a preset threshold are obtained through iterative solution of the marginal energy consumption gain.

[0105] Furthermore, the energy consumption gain solving module 40 is used to perform the following operation steps:

[0106] According to the first initial boundary task and the second initial boundary task that make up the first candidate task adjustment, locate the second spatio-temporal coupled task group and the third spatio-temporal coupled task group in the M spatio-temporal coupled task groups, where the second spatio-temporal coupled task group and the third spatio-temporal coupled task group are processed by the first matching logistics vehicle and the second matching logistics vehicle respectively; perform task exchange migration of the first initial boundary task and the second initial boundary task between the second spatio-temporal coupled task group and the third spatio-temporal coupled task group to obtain a second migrated task group and a third migrated task group; perform logistics energy consumption fitting on the second migrated task group and the third migrated task group respectively based on the first matching logistics vehicle and the second matching logistics vehicle to obtain a first migrated logistics energy consumption and a second migrated logistics energy consumption; extract a first logistics energy consumption baseline and a second logistics energy consumption baseline from the M logistics energy consumption baselines according to the second spatio-temporal coupled task group and the third spatio-temporal coupled task group; if the marginal energy consumption gain obtained by adding the first migrated logistics energy consumption and the second migrated logistics energy consumption is better than the marginal energy consumption gain obtained by adding the first logistics energy consumption baseline and the second logistics energy consumption baseline, then retain the second migrated task group and the third migrated task group; and so on, perform inter-group task migration according to the number of candidate task adjustments, and solve for the marginal energy consumption gain based on the M matching logistics vehicles to obtain the M migrated task groups.

[0107] Furthermore, the hierarchical clustering module 20 is used to perform the following operation steps:

[0108] Enumerate and combine the multiple logistics tasks to be allocated to obtain W pairs of logistics tasks to be allocated; refer to the W pairs of logistics tasks to be allocated and combine the multiple task spatio-temporal attributes to obtain W pairs of task spatio-temporal attributes; calculate the spatio-temporal overlap degree according to the W pairs of task spatio-temporal attributes to obtain W task spatio-temporal overlap degrees; based on the W task spatio-temporal overlap degrees, perform hierarchical clustering on the multiple logistics tasks to be allocated to obtain the M spatio-temporal coupled task groups.

[0109] Furthermore, the hierarchical clustering module 20 is used to perform the following operation steps:

[0110] Decompose the spatio-temporal attributes of the W pairs of tasks to obtain W pairs of logistics start and end points and W pairs of logistics time windows; locally call the average task transportation volume, and calculate the maximum reusable distance based on the average task transportation volume and the W pairs of logistics start and end points, and output W logistics distance overlap degrees; calculate the window intersection percentage of the W pairs of logistics time windows to obtain W time window overlap degrees; pre-configure the logistics weight relationship to perform weighted fusion of the W logistics distance overlap degrees and the W time window overlap degrees to obtain the W task spatio-temporal overlap degrees.

[0111] Furthermore, the hierarchical clustering module 20 is used to perform the following operation steps:

[0112] Based on the W task spatio-temporal overlap degrees and the W logistics tasks to be allocated, construct a symmetric matrix to obtain a task similarity matrix; locally call the logistics load constraints, where the logistics load constraints include the lower limit of logistics quality, the lower limit of logistics volume, and the lower limit of vehicle redundancy; based on the logistics load constraints and multiple load capacity constraint attributes, perform multiple rounds of descending cluster merging on the task similarity matrix, and output the M spatio-temporal coupling task groups.

[0113] Furthermore, the hierarchical clustering module 20 is used to perform the following operation steps:

[0114] Take the lower limit of logistics quality and the lower limit of logistics volume as cluster division conditions, and combine the multiple load capacity constraint attributes to randomly divide task clusters in the task similarity matrix to obtain K random clusters; after calculating the average similarity of tasks between clusters after combining and enumerating the K random clusters, take the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combine the similarity sorting to aggregate the K random clusters into Q initial aggregated clusters; after calculating the average similarity of tasks between clusters after combining and enumerating the Q initial aggregated clusters, take the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combine the similarity sorting to aggregate the Q initial aggregated clusters into F second aggregated clusters; and so on, perform multiple rounds of descending cluster merging based on similarity until the number of clusters is reduced to meet the lower limit of vehicle redundancy, and extract the M spatio-temporal coupling task groups, where K > Q > F > M, and K, Q, F, M are positive integers.

[0115] Through the foregoing detailed description of a logistics vehicle dynamic scheduling method in this specification, those skilled in the art can clearly know a logistics vehicle dynamic scheduling system in this embodiment. Since it corresponds to the method disclosed in the embodiment, it is described relatively simply. For related parts, refer to the description in the method part.

[0116] Example 3 provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, any step of Example 1 is implemented.

[0117] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0118] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A dynamic scheduling method for logistics vehicles, characterized in that, The method includes: Receiving a plurality of logistics tasks to be assigned, where the plurality of logistics tasks to be assigned have a plurality of task spatio-temporal attributes and a plurality of load capacity constraint attribute identifiers; Performing hierarchical clustering on the plurality of logistics tasks to be assigned according to the plurality of task spatio-temporal attributes to obtain M spatio-temporal coupling task groups; Locally invoking the H vehicle attributes of H idle logistics vehicles, and performing logistics energy consumption fitting according to the H vehicle attributes and the plurality of load capacity constraint attributes, so as to screen M matching logistics vehicles corresponding to the M spatio-temporal coupling task groups from the H idle logistics vehicles; Taking the M matching logistics vehicles as constraints, solving the marginal energy consumption gain of the M spatio-temporal coupling task groups to obtain M energy consumption optimization task groups; Extracting the M logistics trajectories of the M energy consumption optimization task groups to construct M job scheduling sequences; Taking the M job scheduling sequences and the M logistics trajectories as M physical distribution dynamic scheduling strategies to drive the M matching logistics vehicles to execute the plurality of logistics tasks to be assigned; Among them, locally invoking the H vehicle attributes of H idle logistics vehicles, and performing logistics energy consumption fitting according to the H vehicle attributes and the plurality of load capacity constraint attributes, so as to screen M matching logistics vehicles corresponding to the M spatio-temporal coupling task groups from the H idle logistics vehicles, the method includes: Decomposing the task spatio-temporal attributes of the first spatio-temporal coupling task group to extract R logistics start and end points of R constituent tasks; Performing shortest logistics route fitting based on the R logistics start and end points to obtain a first logistics trajectory and a first job scheduling sequence; According to the R constituent tasks, calling R load capacities from the plurality of load capacity constraint attributes; Performing global logistics energy consumption simulation on the H idle logistics vehicles according to the H vehicle attributes, the first job scheduling sequence, the first logistics trajectory, and the R load capacities to obtain a first set of logistics energy consumption information, where the first set of logistics energy consumption information includes H first global logistics energy consumptions; And so on, through logistics energy consumption fitting, obtaining M sets of logistics energy consumption information of the M spatio-temporal coupling task groups; Taking energy consumption minimization as the goal, resolving the allocation conflicts of the H idle logistics vehicles according to the M sets of logistics energy consumption information, and outputting the M matching logistics vehicles.

2. The dynamic scheduling method for a logistics vehicle according to claim 1, wherein Taking the M matching logistics vehicles as constraints, solving the marginal energy consumption gain of the M spatio-temporal coupling task groups to obtain M energy consumption optimization task groups, the method includes: According to the M matching logistics vehicles, extracting M logistics energy consumption baselines from the M sets of logistics energy consumption information; Calculating the spatio-temporal overlap degree of the M spatio-temporal coupling task groups according to the plurality of task spatio-temporal attributes to screen M initial boundary tasks; Calculate the task similarity of the M initial boundary tasks according to the multiple load capacity constraint attributes, and combine the M initial boundary tasks according to the calculation results to obtain candidate task adjustment pairs; According to the said inter-group task migration is performed on the M spatio-temporal coupling task groups according to the candidate task adjustment pairs, and marginal energy consumption gain is solved according to the M matching logistics vehicles to obtain M migrated task groups; And so on, after calculating the spatio-temporal overlap degree of the M migration task groups according to the plurality of task spatio-temporal attributes to screen M updated boundary tasks, performing inter-group task migration and marginal energy consumption gain solution to update the M migration task groups until the M energy consumption optimization task groups with an energy consumption gain less than a preset threshold are obtained through iterative solution of the marginal energy consumption gain.

3. The dynamic scheduling method of a logistics vehicle according to claim 2, characterized in that, With the M matching logistics vehicles as constraints, solve the marginal energy consumption gain for the M spatio-temporal coupling task groups to obtain M energy consumption optimization task groups. The method includes: According to the first initial boundary task and the second initial boundary task that constitute the first candidate task adjustment, locate the second spatio-temporal coupling task group and the third spatio-temporal coupling task group in the M spatio-temporal coupling task groups, where the second spatio-temporal coupling task group and the third spatio-temporal coupling task group are processed by the first matching logistics vehicle and the second matching logistics vehicle respectively; Perform task exchange migration of the first initial boundary task and the second initial boundary task between the second spatio-temporal coupling task group and the third spatio-temporal coupling task group to obtain a second migration task group and a third migration task group; According to the first matching logistics vehicle and the second matching logistics vehicle, respectively perform logistics energy consumption fitting on the second migration task group and the third migration task group to obtain a first migration logistics energy consumption and a second migration logistics energy consumption; Extract a first logistics energy consumption baseline and a second logistics energy consumption baseline from the M logistics energy consumption baselines according to the second spatio-temporal coupling task group and the third spatio-temporal coupling task group; If the marginal energy consumption gain obtained by adding the first migration logistics energy consumption and the second migration logistics energy consumption is better than the marginal energy consumption gain obtained by adding the first logistics energy consumption baseline and the second logistics energy consumption baseline, then retain the second migration task group and the third migration task group; And so on, according to the said candidate tasks, inter-group task migration is adjusted, and the marginal energy consumption gain is solved according to the M matching logistics vehicles to obtain the M migration task groups.

4. The dynamic scheduling method for a logistics vehicle according to claim 1, characterized in that, According to the spatio-temporal attributes of the multiple tasks, perform hierarchical clustering on the multiple to-be-allocated logistics tasks to obtain M spatio-temporal coupling task groups. The method includes: Combination-enumerate the multiple to-be-allocated logistics tasks to obtain W pairs of to-be-allocated logistics tasks; Refer to the W pairs of to-be-allocated logistics tasks and combine the spatio-temporal attributes of the multiple tasks to obtain W pairs of task spatio-temporal attributes; Perform spatio-temporal overlap degree calculation according to the W pairs of task spatio-temporal attributes to obtain W task spatio-temporal overlap degrees; According to the W task spatio-temporal overlap degrees, perform hierarchical clustering on the multiple to-be-allocated logistics tasks to obtain the M spatio-temporal coupling task groups.

5. The dynamic scheduling method for a logistics vehicle according to claim 4, wherein, Perform spatio-temporal overlap degree calculation according to the W pairs of task spatio-temporal attributes to obtain W task spatio-temporal overlap degrees. The method includes: Decompose the W pairs of task spatio-temporal attributes to obtain W pairs of logistics starting and ending points and W pairs of logistics time windows; Locally call the task transportation mean value, and perform the maximum reusable distance calculation according to the task transportation mean value and the W pairs of logistics starting and ending points, and output W logistics distance overlap degrees; Perform window intersection percentage calculation on the W pairs of logistics time windows to obtain W time window overlap degrees; Perform weighted fusion of the W logistics distance overlap degrees and the W time window overlap degrees with a pre-configured logistics weight relationship to obtain the W task spatio-temporal overlap degrees.

6. The dynamic scheduling method for a logistics vehicle according to claim 5, characterized in that, According to the W task spatio-temporal overlap degrees, perform hierarchical clustering on the multiple to-be-allocated logistics tasks to obtain the M spatio-temporal coupling task groups. The method includes: Based on the W task spatio-temporal overlap degrees and the W pairs of to-be-allocated logistics tasks, construct a symmetric matrix to obtain a task similarity matrix; Locally call the logistics load constraint, where the logistics load constraint includes a logistics mass lower limit, a logistics volume lower limit, and a vehicle redundancy lower limit; Based on the logistics load constraint and multiple load capacity constraint attributes, perform multiple rounds of descending cluster merging on the task similarity matrix, and output the M spatio-temporal coupling task groups.

7. The dynamic scheduling method for a logistics vehicle according to claim 6, wherein Based on the logistics load constraint and multiple load capacity constraint attributes, perform multiple rounds of descending cluster merging on the task similarity matrix, and output the M spatio-temporal coupling task groups. The method includes: Use the lower limit of logistics quality and the lower limit of logistics volume as cluster division conditions, and combine the multiple load capacity constraint attributes to randomly divide task clusters in the task similarity matrix to obtain K random clusters. After calculating the average similarity of tasks between clusters after combining and enumerating the K random clusters, use the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combine the similarity sorting to aggregate the K random clusters into Q initial aggregated clusters. After calculating the average similarity of tasks between clusters after combining and enumerating the Q initial aggregated clusters, use the lower limit of logistics quality and the lower limit of logistics volume as cluster aggregation conditions, and combine the similarity sorting to aggregate the Q initial aggregated clusters into F second aggregated clusters. And so on, perform multiple rounds of descending cluster merging based on similarity until the number of clusters is reduced to meet the vehicle redundancy lower limit, and extract the M spatio-temporal coupling task groups, where K > Q > F > M, and K, Q, F, and M are positive integers.

8. A dynamic scheduling system for logistics vehicles, characterized in that, A system for implementing the logistics vehicle dynamic scheduling method according to any one of claims 1-7, the system includes: A logistics task receiving module, configured to receive a plurality of to-be-allocated logistics tasks, where the plurality of to-be-allocated logistics tasks have a plurality of task spatio-temporal attributes and a plurality of load capacity constraint attribute identifiers. A hierarchical clustering module, configured to perform hierarchical clustering on the plurality of to-be-allocated logistics tasks according to the plurality of task spatio-temporal attributes to obtain M spatio-temporal coupling task groups. A logistics energy consumption fitting module, configured to locally call the H vehicle attributes of H idle logistics vehicles, and perform logistics energy consumption fitting according to the H vehicle attributes and the multiple load capacity constraint attributes, so as to screen M matching logistics vehicles corresponding to the M spatio-temporal coupling task groups from the H idle logistics vehicles. An energy consumption gain solving module, configured to perform marginal energy consumption gain solving on the M spatio-temporal coupling task groups with the M matching logistics vehicles as constraints to obtain M energy consumption optimization task groups. A scheduling sequence construction module, configured to extract the M logistics trajectories of the M energy consumption optimization task groups to construct M job scheduling sequences. A logistics task execution module, configured to use the M job scheduling sequences and the M logistics trajectories as M physical flow dynamic scheduling strategies to drive the M matching logistics vehicles to execute the plurality of to-be-allocated logistics tasks.

9. A storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, it implements the steps of the logistics vehicle dynamic scheduling method according to any one of claims 1 to 7.

Images