Vehicle scheduling method

Abstract

The invention provides a vehicle scheduling method, which comprises the steps of outputting a target scheduling scheme according to multiple iterative operations; the Yth iteration operation comprises the following steps: obtaining an optimal scheduling scheme corresponding to the (Y-1) th iteration operation; determining the Yth arrangement sequence of the plurality of transport vehicles; for each transportation task, adding a plurality of candidate vehicles meeting a preset task execution condition in the plurality of transportation vehicles into a candidate vehicle pool corresponding to the transportation task according to the Yth arrangement sequence until the number of the candidate vehicles meets a preset adding condition; a target vehicle corresponding to the transportation task is screened out from a candidate vehicle pool corresponding to the transportation task, and an initial scheduling scheme corresponding to the Yth iteration operation is obtained; and comparing the optimal scheduling scheme corresponding to the (Y-1) th iteration operation with the initial scheduling scheme corresponding to the Yth iteration operation to obtain an optimal scheduling scheme corresponding to the Yth iteration operation.

Description

Technical Field

[0001] The present disclosure relates to the technical field of logistics vehicle scheduling, and more particularly to a vehicle scheduling method. Background Art

[0002] With the rapid development of the logistics transportation industry, long-haul trunk transportation and short-haul feeder transportation have been widely applied. In the process of implementing the inventive concept, the inventors found that there are at least the following problems in the related art: due to the uneven distribution of transportation tasks in different regions and times, problems such as vehicle detention and empty vehicle operation may occur, resulting in low vehicle utilization efficiency and increased operating costs. And the related vehicle scheduling methods usually rely on the experience of dispatchers for manual scheduling, which is not only inefficient but also difficult to meet various complex requirements in the scenario of multi-day vehicle scheduling. In addition, the related vehicle scheduling methods are difficult to adapt to the dynamically changing transportation demands, further restricting the optimization space and operating efficiency of the transportation system. Summary of the Invention

[0003] In view of the above problems, the present disclosure provides a vehicle scheduling method, including: outputting a target scheduling plan according to multiple iterative operations, where the target scheduling plan is used to represent the correspondence between multiple transport vehicles and multiple transport tasks; wherein, any Y-th iterative operation includes: obtaining an optimal scheduling plan corresponding to the (Y - 1)-th iterative operation; determining the Y-th arrangement order of multiple transport vehicles; for each transport task, according to the Y-th arrangement order, adding multiple candidate vehicles that meet the preset task execution conditions among the multiple transport vehicles to the candidate vehicle pool corresponding to the transport task until the number of candidate vehicles meets the preset addition condition; for each transport task, screening out the target vehicle corresponding to the transport task from the candidate vehicle pool corresponding to the transport task to obtain an initial scheduling plan corresponding to the Y-th iterative operation; comparing and processing the optimal scheduling plan corresponding to the (Y - 1)-th iterative operation and the initial scheduling plan corresponding to the Y-th iterative operation to obtain an optimal scheduling plan corresponding to the Y-th iterative operation.

[0004] According to an embodiment of the present disclosure, comparing and processing the optimal scheduling plan corresponding to the (Y - 1)-th iterative operation and the initial scheduling plan corresponding to the Y-th iterative operation to obtain an optimal scheduling plan corresponding to the Y-th iterative operation includes: in the case where the scheduling cost of the initial scheduling plan is less than the scheduling cost of the optimal scheduling plan corresponding to the (Y - 1)-th iterative operation, determining whether the initial scheduling plan meets the preset constraint conditions; in the case where the initial scheduling plan meets the preset constraint conditions, taking the initial scheduling plan as the optimal scheduling plan corresponding to the Y-th iterative operation.

[0005] According to an embodiment of the present disclosure, a comparison process is performed on the optimal scheduling plan corresponding to the (Y - 1)-th iteration operation and the initial scheduling plan corresponding to the Y-th iteration operation to obtain the optimal scheduling plan corresponding to the Y-th iteration operation. It further includes: in the case where the initial scheduling plan does not meet the preset constraint conditions, using the optimal scheduling plan corresponding to the (Y - 1)-th iteration operation as the optimal scheduling plan corresponding to the Y-th iteration operation.

[0006] According to an embodiment of the present disclosure, a comparison process is performed on the optimal scheduling plan corresponding to the (Y - 1)-th iteration operation and the initial scheduling plan corresponding to the Y-th iteration operation to obtain the optimal scheduling plan corresponding to the Y-th iteration operation. It further includes: in the case where the scheduling cost of the initial scheduling plan is greater than or equal to the scheduling cost of the optimal scheduling plan corresponding to the (Y - 1)-th iteration operation, using the optimal scheduling plan corresponding to the (Y - 1)-th iteration operation as the optimal scheduling plan corresponding to the Y-th iteration operation.

[0007] According to an embodiment of the present disclosure, the scheduling cost is calculated based on the no-load driving duration, the vehicle waiting duration, and a preset weight.

[0008] According to an embodiment of the present disclosure, the preset constraint conditions include: task feasibility constraint, vehicle operation continuity constraint, and vehicle return constraint.

[0009] According to an embodiment of the present disclosure, screening out the target vehicle corresponding to the transportation task from the candidate vehicle pool corresponding to the transportation task includes: screening out the target vehicle corresponding to the transportation task based on the vehicle's place of origin, the starting place and the ending place of the transportation task, and the vehicle transportation cost.

[0010] According to an embodiment of the present disclosure, screening out the target vehicle corresponding to the transportation task based on the vehicle's place of origin, the starting place and the ending place of the transportation task, and the vehicle transportation cost includes: in the case where there are preferred vehicles among multiple candidate vehicles whose place of origin belongs to the starting place or the ending place of the transportation task, and the number of preferred vehicles is multiple, determining the target vehicle from the preferred vehicles according to the vehicle transportation cost of the preferred vehicles.

[0011] According to an embodiment of the present disclosure, in the case where the number of preferred vehicles is one, using the preferred vehicle as the target vehicle.

[0012] According to an embodiment of the present disclosure, in the case where there are no preferred vehicles among multiple candidate vehicles, determining the target vehicle from the candidate vehicles according to the vehicle transportation cost of the candidate vehicles.

[0013] According to an embodiment of the present disclosure, by introducing a candidate vehicle pool mechanism, the vehicle screening range can be narrowed, the amount of calculation can be reduced, and the algorithm can efficiently meet the requirements of vehicle regression. Through iterative optimization, and the vehicle order and the size of the candidate vehicle pool are randomly changed in each iteration operation, the algorithm has strong search ability, can avoid falling into local optimal solutions, and can find global optimal or approximate optimal scheduling solutions, so that the greedy strategy can better adapt to the requirements of efficient model solving, and finally a more reasonable scheduling solution can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Through the following description of the embodiments of the present disclosure with reference to the drawings, the above content and other objects, features and advantages of the present disclosure will become clearer. In the drawings:

[0015] Figure 1 Schematically shows a flowchart of a vehicle scheduling method according to an embodiment of the present disclosure;

[0016] Figure 2 Schematically shows a flowchart of a vehicle scheduling method according to another embodiment of the present disclosure;

[0017] Figure 3 Schematically shows a schematic diagram of an objective scheduling solution according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for the sake of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, obviously, one or more embodiments can also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present disclosure. The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "including", "comprising" and the like used herein indicate the presence of features, steps, operations and / or components, but do not exclude the presence or addition of one or more other features, steps, operations or components. All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0019] In the process of implementing the inventive concept, the inventors found that there are at least the following problems in the related art: During the execution of trunk and branch line transportation operations, problems such as vehicle detention and empty vehicle operation may occur. For example, after a vehicle arrives at its destination, due to the lack of return cargo, the vehicle needs to wait for a new transportation task, resulting in vehicle idle detention, or the vehicle returns to the departure place or goes to the next destination empty without cargo, causing waste of transport capacity. Related methods that rely on the experience of dispatchers for scheduling are difficult to meet various complex requirements in the scenario of multi-day vehicle scheduling. Various complex requirements include, for example, when planning and arranging the transportation tasks of vehicles across multiple days, it is necessary to reasonably allocate vehicles to ensure traffic flow balance; in addition, the transportation tasks of vehicles also need to be continuously arranged to avoid long-term vehicle idleness, so as to meet the time continuity requirements; when performing multi-day scheduling (for example, it takes multiple days for a transport vehicle to complete all transportation tasks), the vehicle also needs to return to the departure place or a designated location within a certain time for maintenance or reallocation of tasks, that is, the vehicle needs to meet the vehicle return requirements. In addition, related vehicle scheduling methods lack optimization means based on data and algorithms, are difficult to cope with dynamically changing demands, and often only focus on local tasks, lacking global optimization of the entire transportation network. Therefore, they lack globality and scientificity, resulting in low transportation efficiency and being unable to achieve effective optimization to adapt to the continuous changes in transportation demands.

[0020] In view of this, an embodiment of the present disclosure provides a vehicle scheduling method, including operation S110: Output a target scheduling plan according to multiple iterative operations, where the target scheduling plan is used to represent the correspondence between multiple transport vehicles and multiple transport tasks.

[0021] Specifically, the target scheduling plan can represent the transport tasks to be executed by each transport vehicle and the sequence of precedence between the transport tasks. For example, the scheduling plan includes that vehicle A executes task 1, then task 4, and then task 5; vehicle B first executes task 2 and then task 3.

[0022] According to an embodiment of the present disclosure, transport task data, transport vehicle data, and the no-load driving duration between cities can be read. The transport task data includes data such as task serial number, task start city, task end city, task duration, time window information (including the earliest start time and the latest start time of the task), etc. The transport vehicle data includes data such as vehicle serial number, vehicle affiliated city, vehicle departure city, vehicle available duration, etc. The no-load driving duration is used to represent the duration required for a transport vehicle to travel between different cities in an empty state (that is, the state in which the transport vehicle travels without loading goods).

[0023] According to an embodiment of the present disclosure, the transportation task data, transportation vehicle data, and the no-load driving duration between cities can be initialized to obtain an initialized transportation task set, transportation vehicle set, and the no-load driving duration matrix between cities. Multiple iterative operations can be performed based on the initialized data to output a target scheduling scheme. Since the solution time at the same number of iterations increases as the problem scale increases, the maximum number of iterations can be determined according to the quality of the solution and the solution time. For example, the maximum number of iterations can be set to 10,000 times to ensure the solution efficiency, or the maximum number of iterations can be dynamically adjusted according to the problem scale. The value of the maximum number of iterations is not limited herein.

[0024] Figure 1 Schematically shows a flowchart of an arbitrary Y-th iterative operation according to an embodiment of the present disclosure. As Figure 1 shown, the iterative operation of this embodiment includes operation S111 to operation S115.

[0025] In operation S111, obtain the preferred scheduling scheme corresponding to the (Y - 1)-th iterative operation.

[0026] According to an embodiment of the present disclosure, each iterative operation can generate a corresponding initial scheduling scheme. The initial scheduling scheme obtained in each iterative operation can be compared with the preferred scheduling scheme of the previous iterative operation, and the scheduling scheme that meets the conditions is used as the preferred scheduling scheme for the current iterative operation. For example: The scheduling scheme 1 is obtained in the 1st iteration. After the scheduling scheme 2 is obtained in the 2nd iterative operation, if the scheduling scheme 2 does not meet the conditions, the scheduling scheme 2 is discarded, and the scheduling scheme 1 is used as the preferred scheduling scheme corresponding to the 2nd iterative operation; if the scheduling cost of the scheduling scheme 2 is lower than that of the scheduling scheme 1 and the scheduling scheme 2 meets the conditions, the scheduling scheme 2 is used as the preferred scheduling scheme corresponding to the 2nd iterative operation.

[0027] In operation S112, determine the Y-th arrangement order of multiple transportation vehicles.

[0028] According to an embodiment of the present disclosure, each iterative operation can randomly sort multiple transportation vehicles, generate a random sequence with the same number as the number of vehicles, and reorder the vehicle sequence. For example, if there are 5 transportation vehicles, the vehicle sequence generated in the (Y - 1)-th iterative operation is [3, 1, 5, 2, 4], and in the Y-th iterative operation, this vehicle sequence needs to be reordered, and the randomly generated vehicle sequence corresponding to the Y-th iterative operation is vehicle [4, 3, 5, 2, 1]. By randomly sorting multiple transportation vehicles in the transportation vehicle set in each iterative operation, the algorithm can be prevented from falling into a local optimal solution and the possibility of finding a better solution can be increased.

[0029] According to an embodiment of the present disclosure, transportation tasks can be assigned to transportation vehicles through a greedy strategy, which may specifically include operations S113 to S114.

[0030] In operation S113, for each transportation task, according to the Y-th arrangement order, multiple candidate vehicles that meet the preset task execution conditions among multiple transportation vehicles are added to the candidate vehicle pool corresponding to the transportation task until the number of candidate vehicles meets the preset addition condition.

[0031] According to an embodiment of the present disclosure, when assigning transportation vehicles to each transportation task, a candidate vehicle pool corresponding to the transportation task can be constructed first based on data such as the current position of the transportation vehicle, the available time of the vehicle, and the time window.

[0032] Specifically, the conditions for constructing the candidate vehicle pool include: the sum of the no-load driving duration required for the transportation vehicle to travel from the current position to the task starting place and the current available time of the vehicle is less than or equal to the latest start time of the transportation task, and the transportation task has not been assigned to other transportation vehicles. The current available time of the vehicle includes the time when the transportation vehicle can currently start executing the transportation task. For example, it can be the time when the transportation vehicle completes the previous transportation task.

[0033] For example: A certain transportation vehicle completes the previous transportation task at 9:00 am, then its current available time is 9:00. When the sum of the no-load driving duration required for the transportation vehicle to travel from the current position to the task starting city and the current available time of the vehicle is less than or equal to the latest start time of the task, it indicates that the transportation vehicle can reach the starting city of the task and start executing the task within the specified time.

[0034] For example: A certain transportation vehicle is currently in City A, and the task starting city is City B. The no-load driving duration from City A to City B is 2 hours. The current available time of this transportation vehicle is 9:00, and the latest start time of this transportation task is 11:00 am. Since the sum of the no-load driving duration of this transportation vehicle and the current available time of the vehicle is equal to the latest start time of the transportation task, this transportation vehicle can reach the starting city of the task before the latest start time of the task.

[0035] According to an embodiment of the present disclosure, meeting the preset task execution conditions may include that the sum of the no-load driving duration required for the transportation vehicle to travel from the current position to the task starting city and the current available time of the vehicle is less than or equal to the latest start time of the transportation task, and the vehicle transportation cost of the transportation vehicle is relatively low.

[0036] According to an embodiment of the present disclosure, the preset addition condition may include that the number of candidate vehicles reaches the capacity limit of the candidate vehicle pool, or all candidate vehicles that meet the preset task execution conditions among multiple transportation vehicles have been added to the candidate vehicle pool.

[0037] According to an embodiment of the present disclosure, in each iteration operation, the sizes of the candidate vehicle pools corresponding to each transportation task are the same. The size of the candidate vehicle pool can be re-determined when performing different iteration operations. The upper limit of the capacity of the candidate vehicle pool can be the size s of the candidate vehicle pool, which can be dynamically adjusted according to the number of tasks and actual requirements. The value of s is an integer randomly selected between two numbers. For example, if the number of transportation vehicles is 20 and the size range of the candidate vehicle pool is set to [5, 15], the size of the candidate pool randomly generated in the first iteration operation is 8, and the size of the candidate pool randomly generated in the second iteration operation is 15...

[0038] According to an embodiment of the present disclosure, the size of the candidate vehicle pool is regenerated each time an iteration is performed, enabling the algorithm to explore more different scheduling schemes, increasing the diversity of the search, preventing the algorithm from falling into a local optimal solution, and increasing the likelihood of finding a better solution.

[0039] In operation S114, for each transportation task, the target vehicle corresponding to the transportation task is screened out from the candidate vehicle pool corresponding to the transportation task, obtaining the initial scheduling scheme corresponding to the Y-th iteration operation.

[0040] According to an embodiment of the present disclosure, screening out the target vehicle from the candidate vehicle pool corresponding to the transportation task can be, for example, screening out a transportation vehicle whose vehicle origin belongs to the starting place or the ending place of the transportation task and whose scheduling cost is relatively low as the target vehicle, and the target vehicle is used to perform the transportation task.

[0041] For example, when allocating a transportation vehicle to transportation task 1, the starting place of transportation task 1 is city A. The candidate vehicle pool includes transportation vehicle 1 and transportation vehicle 4. The vehicle origin of transportation vehicle 1 is city A, and the scheduling cost of transportation vehicle 1 is relatively low. Then, transportation vehicle 1 is used as the target vehicle, and transportation vehicle 1 is used to perform transportation task 1.

[0042] According to an embodiment of the present disclosure, for the k-th iteration operation, multiple transportation tasks can be sorted in ascending order according to the earliest start time of the transportation tasks to obtain a task list. Target vehicles can be allocated to each transportation vehicle in sequence according to the arrangement order of the multiple transportation tasks in the task list. Specifically, after allocating a target vehicle to the z-th transportation task in the task list, the status information of the z-th transportation task and the target vehicle can be updated. For example, the status of the z-th transportation task can be updated to "allocated". Updating the status information of the target vehicle can include: updating the current location of the target vehicle to the task termination location of the z-th transportation task, updating the available time of the target vehicle to the end time of the z-th transportation task, updating the initial scheduling plan for the k-th iteration operation, and adding the z-th transportation task. Moreover, if the task start location or the task termination location of the z-th transportation task is the vehicle home location, the return flag of the target vehicle is set to "returned". For the (z + 1)-th transportation task in the task list, a target vehicle corresponding to the (z + 1)-th transportation task is determined according to the updated status information of the target vehicle and the status information of other transportation vehicles. Among them, the "returned" return flag indicates that the task start location or the task termination location of the transportation task is the vehicle home location of the transportation vehicle. In the case where the task start location or the task termination location of the transportation task is the vehicle home location of the transportation vehicle, the transportation vehicle can pass through its vehicle home location during the execution of the transportation task. The task start location can include the task start city, and the task termination location can include the task termination city.

[0043] In operation S115, the preferred scheduling plan corresponding to the (Y - 1)-th iteration operation is compared with the initial scheduling plan corresponding to the Y-th iteration operation to obtain the preferred scheduling plan corresponding to the Y-th iteration operation.

[0044] According to an embodiment of the present disclosure, the scheduling costs of the preferred scheduling plan and the initial scheduling plan can be compared, and whether the preferred scheduling plan and the initial scheduling plan meet the preset constraint conditions can also be compared to obtain the preferred scheduling plan corresponding to the Y-th iteration operation.

[0045] According to the embodiments of the present disclosure, by introducing a candidate vehicle pool mechanism, the vehicle screening range can be narrowed, the computational amount can be reduced, and the algorithm can be ensured to efficiently meet the requirements of vehicle regression. Through iterative optimization, and the vehicle order and the candidate vehicle pool size in each iterative operation are randomly changed, the algorithm has strong search ability, can avoid falling into local optimal solutions, and can find a global optimal or approximate optimal scheduling scheme, enabling the greedy strategy to better adapt to the requirements of efficient model solving, and finally obtaining a relatively reasonable scheduling scheme. Moreover, the structure of the entire vehicle scheduling method is clear, the steps are explicit, it has good operability and scalability, is applicable to various trunk transportation scheduling scenarios, and has broad application prospects and practical value. In addition, the above vehicle scheduling method can handle dynamically changing transportation tasks and vehicle information, realize real-time update and optimization of the scheduling scheme, and meet the requirements of high efficiency and flexibility of logistics transportation.

[0046] According to the embodiments of the present disclosure, comparing and processing the preferred scheduling scheme corresponding to the (Y - 1)-th iterative operation with the initial scheduling scheme corresponding to the Y-th iterative operation to obtain the preferred scheduling scheme corresponding to the Y-th iterative operation includes: when the scheduling cost of the initial scheduling scheme is less than the scheduling cost of the preferred scheduling scheme corresponding to the (Y - 1)-th iterative operation, determining whether the initial scheduling scheme meets the preset constraint conditions; when the initial scheduling scheme meets the preset constraint conditions, taking the initial scheduling scheme as the preferred scheduling scheme corresponding to the Y-th iterative operation.

[0047] According to the embodiments of the present disclosure, the goal of multiple iterative operations is to find a scheduling scheme with the lowest scheduling cost and meeting all preset constraint conditions. Therefore, the scheduling costs of the initial scheduling scheme corresponding to the Y-th iterative operation and the preferred scheduling scheme corresponding to the (Y - 1)-th iterative operation can be compared first. When the scheduling cost of the initial scheduling scheme corresponding to the Y-th iterative operation is lower, it indicates that the initial scheduling scheme of the Y-th iterative operation has an advantage in terms of cost.

[0048] Furthermore, it can be verified whether the initial scheduling scheme meets the preset constraint conditions, such as whether it meets the task feasibility constraint, vehicle operation continuity constraint, and vehicle regression constraint. When the above constraint conditions are met, the initial scheduling scheme of the Y-th iterative operation can be updated to the preferred scheduling scheme corresponding to the Y-th iterative operation.

[0049] According to the embodiments of the present disclosure, through multiple iterative operations, and comparing the scheduling costs of the initial scheduling scheme corresponding to each iterative operation with the preferred scheduling scheme corresponding to the previous iterative operation in each iterative operation, and determining whether the scheduling scheme meets the preset constraint conditions, it is possible to finally screen out a scheduling scheme with a relatively low scheduling cost and that can meet all preset constraint conditions as much as possible through multiple iterative operations.

[0050] According to an embodiment of the present disclosure, a comparison process is performed between the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation and the initial scheduling scheme corresponding to the Y-th iteration operation to obtain the preferred scheduling scheme corresponding to the Y-th iteration operation. It further includes: in the case where the initial scheduling scheme does not meet the preset constraint conditions, using the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation as the preferred scheduling scheme corresponding to the Y-th iteration operation.

[0051] According to an embodiment of the present disclosure, since the goal of multiple iteration operations is to find a scheduling scheme with the lowest total cost and meeting all preset constraint conditions, in the case where the initial scheduling scheme does not meet the preset constraint conditions, it indicates that the initial scheduling scheme cannot meet the goal of the above multiple iteration operations. Therefore, the initial scheduling scheme can be discarded, and the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation is used as the preferred scheduling scheme corresponding to the Y-th iteration operation.

[0052] According to an embodiment of the present disclosure, a comparison process is performed between the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation and the initial scheduling scheme corresponding to the Y-th iteration operation to obtain the preferred scheduling scheme corresponding to the Y-th iteration operation. It further includes: in the case where the scheduling cost of the initial scheduling scheme is greater than or equal to the scheduling cost of the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation, using the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation as the preferred scheduling scheme corresponding to the Y-th iteration operation.

[0053] According to an embodiment of the present disclosure, in the case where the scheduling cost of the initial scheduling scheme is greater than or equal to the scheduling cost of the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation, it indicates that the initial scheduling scheme cannot meet the requirement of the lowest scheduling cost in the goal of multiple iteration operations. Therefore, the initial scheduling scheme can be directly discarded, and the preferred scheduling scheme corresponding to the (Y - 1)-th iteration operation is used as the preferred scheduling scheme corresponding to the Y-th iteration operation, without further determining whether the initial scheduling scheme meets the preset constraint conditions.

[0054] According to an embodiment of the present disclosure, directly discard the initial scheduling scheme when it has no cost advantage without performing subsequent verification of preset constraint conditions, avoiding unnecessary calculations, thereby ensuring both the efficiency of the vehicle scheduling method and that the finally output target scheduling scheme is globally optimal or approximately optimal.

[0055] According to an embodiment of the present disclosure, the vehicle scheduling method further includes: calculating the scheduling cost according to the no-load driving duration, the vehicle waiting duration, and a preset weight.

[0056] According to an embodiment of the present disclosure, the total no-load driving duration can be obtained based on the no-load driving durations of each transportation task in the scheduling plan, the total vehicle waiting duration can be obtained based on the vehicle waiting durations of each transportation vehicle in the scheduling plan, and the scheduling cost can be calculated according to the total no-load driving duration, the total vehicle waiting duration, and a preset weight. The preset weight includes a no-load driving weight and a waiting time weight, and is used to balance the importance of the total no-load driving duration and the total vehicle waiting duration. The calculation of the scheduling cost can be shown as the following formula (1) or (2).

[0057] (1)

[0058] (2)

[0059] Wherein, NewCost is the scheduling cost, θ t is the no-load driving weight, θ w is the waiting time weight, totaltravel k is the total no-load driving duration, totalwait k is the total vehicle waiting duration. In the initialized transportation vehicle set K, the transportation vehicle can be represented by the vehicle serial number k.

[0060] According to an embodiment of the present disclosure, the vehicle waiting duration is determined based on the difference between the task execution time and the time when the transportation vehicle arrives at the transportation task execution location, and the no-load driving duration is determined based on the time when the transportation vehicle arrives at the transportation task execution location in an empty load state from the current location of the vehicle.

[0061] Specifically, the task execution time may include the earliest start time of the task, the transportation task execution location may include the starting city of the task, and the vehicle waiting duration may specifically be the non-negative value of the difference between the earliest start time of the task and the time when the transportation vehicle arrives at the starting city of the task.

[0062] According to an embodiment of the present disclosure, reducing the no-load driving duration can reduce fuel consumption and vehicle wear, thereby reducing operating costs. Reducing the vehicle waiting duration can improve the utilization efficiency of the vehicle. By comparing the scheduling costs of the initial scheduling plan and the optimal scheduling plan, a scheduling plan with lower operating costs and higher transportation efficiency can be selected.

[0063] According to an embodiment of the present disclosure, the preset constraint conditions include: task feasibility constraint, vehicle operation continuity constraint, vehicle return constraint.

[0064] According to an embodiment of the present disclosure, the task feasibility constraint can ensure that each transportation task can be reasonably assigned to a transportation vehicle and completed within the permitted time range. The task feasibility constraint may include a time window constraint and a task execution constraint.

[0065] Among them, the time window constraint is used to characterize that the arrival time of the transport vehicle at the starting city of the task must be within the time window allowed for the transport task. For example, the arrival time of the transport vehicle at the starting city of the task is between the earliest start time and the latest start time of the task. If the transport vehicle arrives at the starting city of the task too early, it will result in unnecessary waiting time; if the arrival time is too late, the execution of the task will be delayed. The time window constraint ensures that the task can start within the allowed time range, thereby improving the transport efficiency while ensuring the smooth execution of the transport task.

[0066] The task execution constraint is used to ensure that all transport tasks must be assigned to transport vehicles to ensure that each transport task can be completed and to ensure that there are no missing transport tasks.

[0067] According to an embodiment of the present disclosure, the vehicle operation continuity constraint is used to ensure that the scheduling scheme is coherent in time and space, including the traffic flow balance constraint and the time continuity constraint.

[0068] The traffic flow balance constraint is used to characterize that the task sequence of the transport vehicle must be coherent, and the connection between transport tasks meets the continuity in time and space. After the transport vehicle completes a transport task, there must be enough time to drive to the starting location of the next transport task and start the next transport task on time. The time continuity constraint is used to characterize that the time sequence of the transport vehicle executing the transport task must be continuous, which takes into account the influence of the task duration and the empty driving duration.

[0069] According to an embodiment of the present disclosure, the vehicle return constraint can ensure that the transport vehicle must pass through its vehicle home location at least once during the execution of the transport task. The vehicle home location can be the location where the transport vehicle is registered or operates. For example, if the address of the company to which vehicle 1 belongs is city 1, then city 1 can be the vehicle home location of vehicle 1. By ensuring that the transport vehicle must pass through its vehicle home location at least once during the execution of the transport task, the transport vehicle can be enabled to regularly return to the vehicle home location for maintenance, driver shift change and other operations. To ensure that the vehicle can reasonably arrange the return plan within the scheduling cycle and avoid related problems caused by not returning for a long time.

[0070] According to an embodiment of the present disclosure, since the preset constraint conditions include the task feasibility constraint, the vehicle operation continuity constraint, and the vehicle return constraint, various conditions such as traffic flow balance, time window, time continuity, and vehicle return can be comprehensively considered, thereby generating a more scientific and reasonable scheduling scheme.

[0071] Specifically, the preset constraint conditions are represented by the objective function shown in the following formula (3).

[0072]

[0073] Each constraint condition can be specifically expressed as the following formulas (4) to (11).

[0074]

[0075]

[0076]

[0077] (7)

[0078] (8)

[0079] (9)

[0080]

[0081] (11)

[0082] The meanings of the symbols in the above formulas are shown in Table 1 below.

[0083] Table 1

[0084]

[0085] Among them, the objective function represents minimizing the vehicle usage cost, specifically minimizing the sum of the no-load driving duration and the vehicle waiting duration multiplied by their weight coefficients; Constraint (4) means that each transportation task is completed exactly once by one transportation vehicle; Constraint (5) means that each transportation vehicle must start from the virtual starting point and finally reach the virtual ending point; Constraint (6) is the vehicle flow balance constraint for each transportation task to ensure that the inflow and outflow of transportation vehicles at the task nodes are equal; Constraint (7) means that the transportation vehicle must comply with the time window requirements when performing transportation tasks; Constraint (8) means that the transportation vehicle needs to meet the continuous time requirement when performing consecutive transportation tasks; Constraint (9) means the waiting duration between two consecutive transportation tasks performed by the transportation vehicle, and any waiting duration is greater than or equal to 0; Constraint (10) means that each transportation vehicle must meet the requirement of vehicle return during the cycle of performing transportation tasks; Constraint (11) means the value range of the variable.

[0086] For the vehicle virtual starting point and the virtual ending point , in the model, it ensures the establishment of the vehicle flow balance constraint (6), and the virtual starting point has the function of inputting the starting city of the vehicle and the earliest available time information, and the virtual ending point has the function of outputting and saving the ending city of the vehicle and the ending time information. According to the characteristics of trunk transportation, the virtual ending point in the model does not require the vehicle to return to a specific city, so the virtual ending point of the vehicle is actually the ending city of the last transportation task completed by the vehicle.

[0087] Transport task The duration st i Specifically, it includes the total time of activities such as loading, unloading, transportation, and rest of the task. In general trunk transportation, for a vehicle to complete a one-way transportation task, it usually needs to go through the following three links. First, the vehicle arrives at the starting station for loading, then transports the goods to the destination through trunk transportation, and finally arrives at the destination station for unloading to complete the task. The total time taken for the three links is the task duration.

[0088] Symbol w ijk Indicates the waiting duration between when vehicle k departs from transportation task i and when transportation task j starts. Specifically, after vehicle k completes transportation task i, it departs from its terminal city e i and travels empty to the starting city s of the next task j j and waits for the time a when task j starts j The required duration.

[0089] According to an embodiment of the present disclosure, screening out the target vehicle corresponding to the transportation task from the candidate vehicle pool corresponding to the transportation task includes: screening out the target vehicle corresponding to the transportation task according to the vehicle's place of origin, the starting place and the ending place of the transportation task, and the vehicle transportation cost.

[0090] For example: A transportation vehicle whose place of origin belongs to the starting place or the ending place of the transportation task and whose vehicle transportation cost is relatively low can be used as the target vehicle.

[0091] According to an embodiment of the present disclosure, the vehicle's place of origin of the candidate vehicle can be matched with the starting place and the ending place of the transportation task to determine whether the vehicle's place of origin of the candidate vehicle belongs to the starting place and the ending place of the transportation task. Further, when there is an unreturned vehicle among the candidate vehicles and the vehicle's place of origin of this vehicle is the starting place or the ending place of the transportation task, this vehicle can be preferentially used as the target vehicle. In the case where multiple candidate vehicles have not returned and the vehicle's place of origin is the starting place or the ending place of the transportation task, the vehicle with the lowest vehicle transportation cost can be used as the target vehicle. And in the case where multiple candidate vehicles have not returned, the vehicle's place of origin is the starting place or the ending place of the transportation task, and the vehicle transportation costs are the same, the first-come, first-served (FCFS) strategy is adopted, and the transportation vehicle that arrives at the starting place of the task earliest is used as the target vehicle to ensure the fairness and transparency of vehicle scheduling.

[0092] According to an embodiment of the present disclosure, specifically, screening out a target vehicle corresponding to a transportation task based on the vehicle's place of registration of the candidate vehicle, the starting place and the ending place of the transportation task, and the vehicle transportation cost includes: in the case where there are preferred vehicles among multiple candidate vehicles whose place of registration belongs to the starting place or the ending place of the transportation task, and the number of preferred vehicles is multiple, determining the target vehicle from the preferred vehicles according to the vehicle transportation cost of the preferred vehicles.

[0093] According to an embodiment of the present disclosure, in the case where the place of registration of a vehicle belongs to the starting place or the ending place of a transportation task, it can be ensured that the transportation vehicle can return to the place of registration of the vehicle at least once during the execution of the transportation task for operations such as maintenance and driver shift change, so that vehicle resources can be managed and allocated more effectively. In the case where there are multiple preferred vehicles, the preferred vehicle with the lowest vehicle transportation cost can be selected as the target vehicle. The obtained target vehicle can not only meet the requirement that the place of registration of the vehicle belongs to the starting place or the ending place of the transportation task, but also has a relatively low vehicle transportation cost.

[0094] According to an embodiment of the present disclosure, the above-mentioned screening out of a target vehicle corresponding to a transportation task based on the vehicle's place of registration of the candidate vehicle, the starting place and the ending place of the transportation task, and the vehicle transportation cost further includes: in the case where the number of preferred vehicles is one, taking the preferred vehicle as the target vehicle.

[0095] According to an embodiment of the present disclosure, in the case where there is only one preferred vehicle, the preferred vehicle can be directly taken as the target vehicle even if the cost of the preferred vehicle is higher than that of other candidate vehicles.

[0096] According to an embodiment of the present disclosure, the above-mentioned screening out of a target vehicle corresponding to a transportation task based on the vehicle's place of registration of the candidate vehicle, the starting place and the ending place of the transportation task, and the vehicle transportation cost further includes: in the case where there are no preferred vehicles among multiple candidate vehicles, determining the target vehicle from the candidate vehicles according to the vehicle transportation cost of the candidate vehicles.

[0097] According to an embodiment of the present disclosure, in the case where there are no preferred vehicles, the vehicle with the lowest vehicle transportation cost is preferentially selected as the target vehicle. By selecting the vehicle with the lowest cost for transportation, the transportation cost can be reduced, transportation resources can be reasonably allocated, resource waste can be avoided, and the vehicle use efficiency can be improved.

[0098] According to an embodiment of the present disclosure, the vehicle transportation cost can be expressed, for example, by the following formula (12) or (13).

[0099] (12)

[0100] (13)

[0101] Among them, cost is the scheduling cost, θ t is the no-load driving weight, θ w is the waiting time weight, is the no-load driving duration, and wait_time is the vehicle waiting duration.

[0102] Among them, the vehicle waiting duration can be calculated by the following formula (14).

[0103]

[0104] Among them, the vehicle arrival time can be calculated by the following formula (15).

[0105]

[0106] According to the embodiments of the present disclosure, by minimizing the no-load driving duration, the vehicle itinerary can be arranged more efficiently, avoiding long-time driving of transport vehicles without load, thereby reducing the ineffective driving mileage of the vehicles, and also reducing fuel consumption and operating costs. By considering the available time, it can be ensured that the transport vehicle can be immediately put into the task after being ready, avoiding the situation of vehicle idle waiting, making the operation time of the transport vehicle more compact, and further improving the utilization rate of the transport vehicle.

[0107] Figure 2 Schematically shows a flowchart of a vehicle scheduling method according to another embodiment of the present disclosure. As Figure 2 shown, the vehicle scheduling method of this embodiment includes operations S210 to S280.

[0108] In operation S210, data reading and initialization.

[0109] According to the embodiments of the present disclosure, the transport task data, transport vehicle data, and no-load driving duration between cities can be read. A transport task data model is established according to the transport task data, and a vehicle data model is constructed according to the transport vehicle data. According to the cities involved in the transport task and the no-load driving duration between cities, a driving duration matrix between cities is constructed.

[0110] Among them, the transport task data model includes the departure time, origin city, destination city of the transport task, and the duration of the transport task. Specifically, the transport task departure time includes the earliest start time a i and the latest start time b i , if the earliest start time is equal to the latest start time, that is, a i =b i , it means that the transport task must start at a certain moment. The duration st iIt includes the total duration required for in-transit transportation, cargo loading and unloading, and downtime. The origin city represents the city from which the transport vehicle departs to perform the transport task, and the destination city represents the city where the transport task is destined. Labels can be set for the transport task and the only city involved respectively.

[0111] The vehicle data model includes the home city, origin city, and earliest available time of the transport vehicle. The origin city of the transport vehicle represents the city where the transport vehicle is located before performing the transport task. The earliest available time represents the time when the vehicle is ready before performing the task.

[0112] Among them, for the matrix of no-load driving durations between cities, the no-load driving duration t between any two cities ij represents the driving time required for the transport vehicle to travel from city i to city j without loading goods ( ), if i = j, then t ij = 0. The no-load driving durations can be stored in matrix form.

[0113] According to the embodiments of the present disclosure, the transport task data model, vehicle data model, and matrix of driving durations between cities can be initialized to obtain an initialized transport task set, an initialized vehicle set, and a matrix of driving durations.

[0114] For example: The initialized transport task set is used to represent each transport task as a task object containing the task serial number i, origin city s i , destination city e i , task duration st i , earliest start time a of the transport task i and latest start time b i , and sorted in ascending order by the earliest start time a i . The initialized transport vehicle set K can represent each transport vehicle as: a vehicle object containing the vehicle serial number k, home city h k , current city d k , current available time t k , empty task sequence , return flag , total no-load driving duration and total vehicle waiting duration attributes, where the current city of the vehicle is the origin city of the vehicle , and the current available time of the vehicle is the earliest available time of the vehicle . A search structure for no-load driving durations between cities can also be constructed through initialization processing to quickly obtain the no-load driving duration t between any two cities during vehicle scheduling ij . The total cost of the scheduling plan can be initialized to infinity, that is , initialize the iteration counter , and set the maximum number of iterations N.

[0115] In operation S220, determine an initial scheduling plan through a greedy strategy.

[0116] According to an embodiment of the present disclosure, determining an initial scheduling plan through a greedy strategy may include adding a plurality of candidate vehicles that meet the preset task execution conditions to a candidate vehicle pool corresponding to the transportation task, and screening out target vehicles corresponding to the transportation task from the candidate vehicle pool, so as to allocate a transportation vehicle to each transportation task in turn.

[0117] First, the first transportation task in the transportation task set can be extracted, and a candidate vehicle pool can be generated according to conditions such as the current location, available time, and task time window requirements of the transportation vehicle, and the vehicle transportation cost of the transportation vehicle for executing the transportation task can be calculated, and the transportation vehicles can be sorted in ascending order according to the vehicle transportation cost.

[0118] Wherein: The transportation vehicle k departs from the current city to the starting city of the transportation task of needs to be before the latest start time b of the transportation task i , that is .

[0119] The vehicle waiting duration can specifically be the non - negative value of the difference between the earliest start time of the transportation task and the time when the transportation vehicle arrives at the task starting city, and can be expressed as . When the time when the transportation vehicle arrives at the starting city s of transportation task i i has not reached the start time window of the task, that is , the vehicle waiting duration is equal to ; when the time when the vehicle arrives at the starting city s of transportation task i i has already been within the task start time window, that is , at this time the vehicle waiting duration is equal to 0.

[0120] According to an embodiment of the present disclosure, candidate vehicles that meet the preset task execution conditions can be added to the candidate vehicle pool corresponding to the transportation task according to the size s of the candidate vehicle pool. In the case where there are no candidate vehicles that meet the preset task execution conditions, the transportation tasks for which the target vehicles have been determined are removed from the transportation task set. The range of s is dynamically adjusted according to the number of transportation tasks and the actual transportation requirements, and the range of s is an integer randomly selected between two positive integers.

[0121] Furthermore, the target vehicle can be determined from the candidate vehicle pool. Specifically, when there is a transportation vehicle k with a regression flag in the candidate vehicle pool, and the starting city s of the transportation task ​i or the destination city e i is the home city h of the transport vehicle k , that is , then the transport task i is preferentially assigned to the transport vehicle k. In the absence of the above vehicles, a candidate vehicle with a lower cost can be preferentially used as the target vehicle.

[0122] According to an embodiment of the present disclosure, after the target vehicle is assigned to the transport task, the status information of the target vehicle can be updated, specifically expressed as: the current position of the target vehicle , available time , task sequence , return flag , total empty driving duration , total vehicle waiting duration And the transport task to which the target vehicle has been assigned can be removed from the transport task set. According to the updated status information of the target vehicle and the status information of other transport vehicles in the transport vehicle set, target vehicles are assigned to other transport tasks in the transport task set until the transport task set is empty, and an initial scheduling plan is obtained.

[0123] In operation S230, the scheduling cost of the initial scheduling plan is calculated.

[0124] In operation S240, it is determined whether the scheduling cost of the initial scheduling plan is less than the current preferred scheduling plan. If the scheduling cost of the initial scheduling plan is greater than or equal to the current preferred scheduling plan, operation S270 is entered.

[0125] In operation S250, when the scheduling cost of the initial scheduling plan is less than the current preferred scheduling plan, it is determined whether the initial scheduling plan meets the preset constraint conditions. If the initial scheduling plan does not meet the preset constraint conditions, operation S270 is entered.

[0126] In operation S260, when the initial scheduling plan meets the preset constraint conditions, the current preferred scheduling plan is updated.

[0127] According to an embodiment of the present disclosure, it can be further determined whether all transport tasks in the transport task set have been assigned to vehicles, and it is determined that the return flags of all vehicles are 1, that is . If the initial scheduling plan meets the conditions, the current preferred scheduling plan is updated, and the scheduling cost of the current preferred scheduling plan is updated, which can be specifically expressed as . Among them, the current preferred scheduling plan may include the preferred scheduling plan corresponding to the current iterative operation.

[0128] In operation S270, it is determined whether the maximum number of iterations has been reached.

[0129] According to an embodiment of the present disclosure, when the maximum number of iterations is not reached, randomly determine the sorting of the transport vehicles, reset the transport tasks and the status information of the transport vehicles, and re-execute operation S220.

[0130] In operation S280, when the maximum number of iterations is reached, output the target scheduling plan.

[0131] According to an embodiment of the present disclosure, if the current number of iterations m has reached the preset maximum number of iterations N (by default, N = 10,000, which can be dynamically adjusted according to the problem scale), that is, N ≤ m, then terminate the iterative operation and output the target scheduling plan.

[0132] According to an embodiment of the present disclosure, when the maximum number of iterations is not reached, increment the iteration counter by one, that is , re-execute operation S220.

[0133] According to an embodiment of the present disclosure, resetting the transport tasks may include randomly sorting the set of transport vehicles K to obtain a new vehicle list K new , update the set of transport vehicles , reset the current positions, available times, and task sequences of all transport vehicles to their initial states. And reset the set of transport tasks to the unassigned state. According to the reset transport tasks and the status information of the transport vehicles, re-execute operation S220.

[0134] According to an embodiment of the present disclosure, through the above operations S210 to S290, the sum of the no-load driving duration and the waiting duration of all transport vehicles can be minimized as the objective function, and the behavior of the transport vehicles visiting their home cities (i.e., the station cities) during the multi-day execution of transport tasks is comprehensively considered, thereby constructing a multi-day scheduling problem model for trunk transport vehicles considering station visits, so that the vehicles can return to their home cities at a certain time point during the process of completing transport tasks (this process can be called a station).

[0135] According to an embodiment of the present disclosure, by considering the behavior of the transport vehicles visiting their home cities during the multi-day execution of transport tasks, the transport vehicles can return to their home cities during the execution of transport tasks, thereby enabling reasonable allocation and adjustment of transport capacity, avoiding over-concentration of transport vehicles in certain areas, and thus maintaining transport capacity balance. And it is convenient for the transport vehicles to perform necessary maintenance to ensure the normal operation of the vehicles. By comprehensively considering the influence of the no-load driving duration and the waiting duration, the use efficiency of the transport vehicles can be improved, the no-load driving duration and the vehicle waiting duration of the vehicles can be reduced, and thus the operating cost can be reduced.

[0136] According to the embodiments of the present disclosure, by adopting an improved greedy strategy, a candidate vehicle pool is set up, and a random selection mechanism for the size of the candidate vehicle pool and the arrangement order of transportation vehicles is introduced. Combining various constraints such as the current position of the vehicle, available time, and task time window, a relatively reasonable transportation vehicle can be dynamically allocated for each transportation task. Further, by verifying the scheduling scheme based on the scheduling cost and preset constraint conditions, and through multiple iterations of optimization, the phenomenon of empty vehicle running and detention can be reduced, the transportation efficiency can be improved, and the global cost optimization and transport capacity balance can be achieved. Thus, a scientific, reasonable, and low-cost target scheduling scheme can be quickly generated, the utilization rate of transportation vehicles can be improved, and the operation cost of enterprises can be reduced.

[0137] The following further illustrates the present disclosure through embodiments and related test experiments. In the following detailed description, for the sake of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, obviously, one or more embodiments can also be implemented without these specific details. Moreover, without conflict, the details in the following embodiments can be combined arbitrarily into other feasible embodiments.

[0138] Embodiment 1:

[0139] Within 7 days, 24 transportation tasks need to be completed, involving 6 transportation vehicles. For the sake of convenience of explanation, the no-load driving weight θ t is set to 1, the waiting time weight θ w is set to 2, and the size of the candidate vehicle pool is a random integer within the range of [1, 3]. And for the sake of simplifying the explanation, the maximum number of iterations is set to 4. All transportation tasks and transportation vehicles are represented by task numbers and vehicle numbers. It should be noted that the above preset weights, maximum number of iterations, and size of the candidate vehicle pool are only for illustrative purposes, and the values of the preset weights, maximum number of iterations, and size of the candidate vehicle pool are not limited.

[0140] The vehicle scheduling method of the embodiments of the present disclosure may include steps S1 to S5.

[0141] In step S1, the transportation task data, transportation vehicle data, and the no-load driving duration matrix between cities are read and initialized. The transportation task data is shown in Table 2, the transportation vehicle data is shown in Table 3, and the no-load driving duration matrix between cities is shown in Table 4. Among them, the time-related data is in hours, and the earliest start time and the latest start time represent the time within 168 hours of 7 days. For example, the earliest start time 22 represents 22:00 on the first day, and the latest start time 144 represents 0:00 on the 7th day.

[0142] Table 2

[0143]

[0144] Table 3

[0145]

[0146] Table 4

[0147]

[0148] In step S2, the first iterative operation is performed, including allocating transportation tasks to transportation vehicles by using a greedy strategy.

[0149] According to an embodiment of the present disclosure, all transportation tasks can be sorted in ascending order according to the earliest start time to obtain a task list, so as to preferentially process transportation tasks with earlier start times. The task list after ascending sorting can be expressed as the following formula (16). The numbers in the following task list are the task numbers of transportation tasks.

[0150] [1, 3, 8, 18, 20, 6, 13, 15, 23, 9, 21, 4, 11, 16, 24, 2, 7, 14, 19, 5, 10, 12, 17, 22] (16)

[0151] According to an embodiment of the present disclosure, for the first iterative operation, the arrangement order of transportation vehicles can be randomly shuffled to generate a vehicle list corresponding to the first iterative operation. The vehicle list can be expressed as: [3, 5, 1, 4, 2, 6]. The numbers in the above vehicle list are the vehicle numbers of transportation vehicles. By randomly shuffling the arrangement order of transportation vehicles during each iterative operation, the diversity of the search solution space can be increased.

[0152] Transportation vehicles are allocated to each transportation task in sequence according to the sorting of transportation tasks in the task list, including: for each transportation task, calculating the driving time of all transportation vehicles in the vehicle list from the current position to the starting city of the task, and determining whether the transportation vehicle meets the task time window constraint. When the transportation vehicle meets the task time window constraint and the transportation cost is relatively low, it is added to the candidate vehicle pool until the number of vehicles in the candidate vehicle pool reaches the upper limit of the capacity of the candidate vehicle pool or there are no more transportation vehicles that meet the conditions. Among them, the task time window constraint can include that the no-load driving duration of the transportation vehicle from the current position to the starting city of the task plus the current available time of the vehicle does not exceed the latest start time of the task.

[0153] For example: the size of the candidate vehicle pool in the first iterative operation is 2, that is, the upper limit of the capacity of the candidate vehicle pool is 2. For the first transportation task in the task list, that is, task 1, vehicles 2, 5, and 4 in the vehicle list meet the task time window constraint of task 1.

[0154] Among them, as can be seen from Tables 2 to 4, the no-load driving duration of vehicle 2 from the current location city 1 to the starting city city 1 of task 1 is 0. The arrival time of vehicle 2 at the starting city of task 1 is 0 + 10 = 10. The waiting time of vehicle 2 for task 1 is max(0, 22 - 10) = 12. The calculated transportation cost of vehicle 2 is 0 + 2×[22 - (0 + 10)] = 24. Correspondingly, the transportation cost of vehicle 5 is 19.3 + 2×[22 - (0 + 19.3)] = 24.7; the transportation cost of vehicle 4 is 15.3 + 2×[22 - (0 + 15.3)] = 28.7. Since vehicle 2 and vehicle 5 meet the task time window constraint of task 1 and have lower transportation costs, and the upper limit of the capacity of the candidate vehicle pool is 2, vehicle 2 and vehicle 5 will be added to the candidate vehicle pool. The candidate vehicle pool for task 1 can be expressed as [2:24, 5:24.7].

[0155] For vehicles 2 and 5 in the candidate vehicle pool, vehicle 5 has not returned and the vehicle home location of vehicle 5 does not belong to the starting place or ending place of task 1. Therefore, the vehicle 2 with a lower cost is selected as the target vehicle for task 1, and task 1 is executed by vehicle 2.

[0156] According to an embodiment of the present disclosure, after allocating the target task for a certain task in the task list, the status of the task and the target vehicle is updated. For example: Task 1 is marked as allocated, and the status information of vehicle 2 is updated according to Tables 2 to 4, including updating the current location of vehicle 2 to the ending city 2 of task 1; the available time of vehicle 2 is the ending time of task 1, where the ending time of task 1 is 36; the task numbers of the transportation tasks to be executed by vehicle 2 include task 1, and the vehicle return mark is updated to "returned".

[0157] Further, according to the information of the updated target vehicle, a transportation vehicle is allocated for the next transportation task in the task list. For example, a target vehicle is allocated for task 3. Among them, the candidate vehicle pool for task 3 can be expressed as [5:24.7, 4:28.7], indicating that the candidate vehicle pool includes vehicle 5 with a transportation cost of 24.7 and vehicle 4 with a transportation cost of 28.7. Since vehicle 5 has not returned and the ending place of task 3 is the vehicle home location of vehicle 5, task 3 is allocated to vehicle 5. Correspondingly, the status of task 3 is updated to allocated, and the status information of vehicle 5 is updated.

[0158] According to an embodiment of the present disclosure, for the first iteration operation, transportation vehicles can be allocated for all transportation tasks in the task list to obtain an initial scheduling plan corresponding to the first iteration operation.

[0159] In step S3, the scheduling cost of the initial scheduling plan is calculated according to the total no-load driving duration and total vehicle waiting duration of each transportation vehicle.

[0160] For example, the initial scheduling plan may include: the task sequence of vehicle 3 is [8, 2, 12, 22]; the task sequence of vehicle 5 is [3, 6, 21, 24, 14]; the task sequence of vehicle 1 is [18, 15, 4, 7, 10]; the task sequence of vehicle 4 is [13, 16, 19, 5]; the task sequence of vehicle 2 is [1]; the task sequence of vehicle 6 is [20, 23, 9, 11, 17]. The scheduling cost of this initial scheduling plan is calculated to be 571.8 according to the following formula (17).

[0161] (17)

[0162] In step S4, verify the initial scheduling plan for the first iterative operation. Since it is the first iterative operation and there is no optimal scheduling plan for the previous iterative operation, verifying the initial scheduling plan for the first iterative operation may include verifying whether the initial scheduling plan meets the preset constraint conditions. If the preset constraint conditions are met, update the optimal scheduling plan and use the initial scheduling plan for the first iterative operation as the optimal scheduling plan, where the optimal scheduling plan is 571.8.

[0163] In step S5, continue to perform multiple iterative operations.

[0164] For the second iterative operation: At the second iterative operation, shuffle the order of the transport vehicles, regenerate the vehicle list, and re-determine the size of the candidate vehicle pool. For example, the vehicle list corresponding to the second iterative operation is [6, 2, 4, 1, 5, 3], and the size of the candidate vehicle pool is 2. The initial scheduling plan obtained from the second iterative operation may include:

[0165] The task sequence of vehicle 6 is [20, 23, 9, 11]; the task sequence of vehicle 2 is [1, 2, 12, 22]; the task sequence of vehicle 4 is

[14] ; the task sequence of vehicle 1 is [18, 15, 4, 7, 10]; the task sequence of vehicle 5 is [3, 6, 21, 24, 17]; the task sequence of vehicle 3 is [8, 13, 16, 19, 15].

[0166] Furthermore, the scheduling cost of the initial scheduling plan for the second iterative operation is calculated to be 730.7, which is greater than the scheduling cost of the optimal scheduling plan, 571.8. Therefore, discard the initial scheduling plan for the second iterative operation and directly proceed to the third iterative operation.

[0167] For the third iterative operation: Shuffle the order of the transport vehicles again, regenerate the vehicle list, and re-determine the size of the candidate vehicle pool. The new vehicle list obtained is [6, 1, 4, 5, 3, 2], and the size of the candidate vehicle pool is 3. The initial scheduling plan obtained from the third iterative operation may include:

[0168] Vehicle 6 has no transportation task to execute. The task sequence of vehicle 1 is [18, 15, 4, 7, 10], the task sequence of vehicle 4 is [20, 23, 9, 11, 17], the task sequence of vehicle 5 is [3, 6, 21, 24, 14], the task sequence of vehicle 3 is [8, 13, 16, 19, 5], and the task sequence of vehicle 2 is [1, 2, 12, 22].

[0169] Furthermore, the scheduling cost of the initial scheduling plan for the 3rd iteration operation is 539.5, which is lower than the scheduling cost of 571.8 of the preferred scheduling plan corresponding to the 2nd iteration operation. Also, the initial scheduling plan for the 3rd iteration operation meets all the preset constraint conditions. Therefore, the preferred scheduling plan is updated, and the initial scheduling plan for the 3rd iteration operation is used as the preferred scheduling plan corresponding to the 3rd iteration operation. Thus, the scheduling cost of the current preferred scheduling plan is 539.5.

[0170] For the 4th iteration operation: Shuffle the order of the transportation vehicles, regenerate the vehicle list, and re-determine the size of the candidate vehicle pool. The new vehicle list obtained is [5, 1, 3, 2, 4, 6], and the size of the candidate vehicle pool is 3. The initial scheduling plan obtained for the 4th iteration operation includes, for example:

[0171] The task sequence of vehicle 5 is [3, 6, 21, 24, 14], the task sequence of vehicle 1 is [18, 15, 4, 7, 10], the task sequence of vehicle 3 is [8, 13, 16, 19, 5], the task sequence of vehicle 2 is [1, 2, 12, 22], vehicle 4 has no transportation task to execute, and the task sequence of vehicle 6 is [20, 23, 9, 11, 17]. The scheduling cost of the current initial scheduling plan is 517, which is lower than the scheduling cost of 539.5 of the preferred scheduling plan. Then, continue to verify whether the current initial scheduling plan meets the preset constraint conditions. It is found that vehicle 4 does not meet the vehicle return constraint. Therefore, the preferred scheduling plan for the 3rd iteration operation is used as the preferred scheduling plan corresponding to the 4th iteration operation.

[0172] According to the embodiments of the present disclosure, since the current iteration count has reached the maximum iteration count, the iteration operation is stopped, and the target scheduling plan is output, where the scheduling cost of the target scheduling plan is 539.5, and the target scheduling plan is the initial scheduling plan for the 3rd iteration operation.

[0173] Figure 3 A schematic diagram showing the target scheduling plan according to the embodiments of the present disclosure is schematically illustrated.

[0174] As Figure 3As shown in the figure, in the target scheduling plan, vehicle 6 has no transportation task to execute, so it stays in its home city 6. Vehicle 1 first executes task 18, then task 15... and finally task 10. Vehicle 1 departs from city 1, first drives to city 6, then to city 1... and finally arrives at city 5. Among them, the home city of vehicle 1 is city 1. In this target scheduling plan, vehicle 1 can return to city 1 multiple times during the execution of transportation tasks.

[0175] In the above specific embodiments, the purpose, technical solution and beneficial effects of the present invention have been further described in detail. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A vehicle dispatching method, characterized in that: The method comprises: Outputting a target scheduling scheme according to multiple iterative operations, wherein the target scheduling scheme is used to characterize the corresponding relationship between multiple transport vehicles and multiple transport tasks; Among them, any Y-th iteration operation includes: Obtaining a preferred scheduling solution corresponding to the Y-1th iteration operation; Determining a Yth arrangement order of the plurality of transport vehicles; For each of the transport tasks, according to the Yth arrangement order, multiple candidate vehicles among the multiple transport vehicles that meet the preset task execution conditions are added to the candidate vehicle pool corresponding to the transport task, until the number of the candidate vehicles meets the preset addition condition; For each of the transport tasks, a target vehicle corresponding to the transport task is selected from the candidate vehicle pool corresponding to the transport task, and an initial scheduling scheme corresponding to the Yth iterative operation is obtained; The preferred scheduling scheme corresponding to the Y-1th iteration operation is compared with the initial scheduling scheme corresponding to the Yth iteration operation to obtain the preferred scheduling scheme corresponding to the Yth iteration operation.

2. The method according to claim 1, characterized in that The comparing the preferred scheduling scheme corresponding to the Y-1th iteration operation with the initial scheduling scheme corresponding to the Yth iteration operation to obtain the preferred scheduling scheme corresponding to the Yth iteration operation includes: When the scheduling cost of the initial scheduling solution is less than the scheduling cost of the preferred scheduling solution corresponding to the Y-1th iterative operation, determining whether the initial scheduling solution satisfies a preset constraint condition; In the case where the initial scheduling scheme satisfies the preset constraint condition, the initial scheduling scheme is used as the preferred scheduling scheme corresponding to the Yth iteration operation.

3. The method according to claim 2, characterized in that The comparing the preferred scheduling scheme corresponding to the Y-1th iteration operation with the initial scheduling scheme corresponding to the Yth iteration operation to obtain the preferred scheduling scheme corresponding to the Yth iteration operation further includes: In a case where the initial scheduling scheme does not satisfy the preset constraint condition, the preferred scheduling scheme corresponding to the Y-1th iteration operation is used as the preferred scheduling scheme corresponding to the Yth iteration operation.

4. The method according to claim 2, characterized in that: The comparing the preferred scheduling scheme corresponding to the Y-1th iteration operation with the initial scheduling scheme corresponding to the Yth iteration operation to obtain the preferred scheduling scheme corresponding to the Yth iteration operation further includes: When the scheduling cost of the initial scheduling scheme is greater than or equal to the scheduling cost of the preferred scheduling scheme corresponding to the Y-1th iterative operation, the preferred scheduling scheme corresponding to the Y-1th iterative operation is used as the preferred scheduling scheme corresponding to the Yth iterative operation.

5. The method according to claim 2, characterized in that: The method further comprises: The dispatch cost is calculated based on the idle driving time, the vehicle waiting time and the preset weight.

6. The method according to claim 2, characterized in that The preset constraints include: mission feasibility constraints, vehicle operation continuity constraints, and vehicle regression constraints.

7. The method according to claim 1, characterized in that The step of selecting a target vehicle corresponding to the transport task from the candidate vehicle pool corresponding to the transport task comprises: The target vehicle corresponding to the transport task is screened out according to the vehicle affiliation of the candidate vehicle, the task starting point and the task ending point of the transport task, and the vehicle transport cost.

8. The method according to claim 7, characterized in that The step of screening out the target vehicle corresponding to the transport task according to the vehicle location of the candidate vehicle, the task starting location and the task ending location of the transport task, and the vehicle transport cost comprises: When there is a preferred vehicle among the multiple candidate vehicles whose vehicle location belongs to the task starting point or the task ending point of the transportation task, and there are multiple preferred vehicles, the target vehicle is determined from the preferred vehicles based on the vehicle transportation cost of the preferred vehicle.

9. The method according to claim 8, characterized in that The method further comprises: When the number of the preferred vehicles is one, the preferred vehicle is used as the target vehicle.

10. The method according to claim 8, characterized in that The method further comprises: In a case where the preferred vehicle does not exist among the plurality of candidate vehicles, the target vehicle is determined from the candidate vehicles based on the vehicle transportation costs of the candidate vehicles.

Images