A scheduling method incorporating event priority
By combining an event priority-based scheduling method with adaptive adjustment of the time step and sparse network construction, the contradiction between the accuracy of high-priority tasks and the computational efficiency of large-scale scenarios in the existing water truck scheduling algorithm is resolved, and efficient and accurate scheduling scheme generation is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI AIRPORT AUTHORITY
- Filing Date
- 2026-04-01
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies cannot find a balance between ensuring the scheduling accuracy of high-priority tasks and the computational efficiency of large-scale scenarios, resulting in a contradiction between computational accuracy and efficiency in existing water truck scheduling algorithms.
A scheduling method combining event priorities is adopted, which adaptively adjusts the time step size. High-priority tasks use fine-grained step sizes, while non-critical time periods use coarse-grained step sizes. Combined with sparse network construction and relaxation optimization model, efficient scheduling is achieved.
It significantly improves the scheduling accuracy of high-priority tasks, reduces computational redundancy, reduces decision variables, improves algorithm solution efficiency, and adapts to large-scale real-time scheduling requirements.
Smart Images
Figure CN122175290A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of optimizing the scheduling algorithm of airport fresh water trucks, and particularly relates to a scheduling method combining event priorities. Background Art
[0002] The scheduling of airport fresh water trucks is a key ground support link to ensure the normal operation of flights, and the rationality of its scheduling plan directly affects the airport operation efficiency and flight punctuality rate. Currently, the industry generally uses the spatio-temporal network model to solve the vehicle routing problem with time windows (VRPTW) for fresh water trucks. This model刻画s the spatio-temporal association of vehicle driving and operation by discretizing the time dimension, and is the core technical means to achieve the quantitative optimization of the scheduling plan.
[0003] However, in practical applications, when using the spatio-temporal network model to solve the VRPTW problem, there is always an inherent contradiction between computational efficiency and path planning accuracy, which is also a common technical pain point of this type of scheduling algorithm: If a coarse-grained time step is used to discretize the time dimension, although the network scale can be controlled and the calculation speed can be improved, it will lead to distortion of the task time window constraint description, especially unable to accurately ensure the punctual execution of high-priority tasks (such as the guarantee of key flights); If a globally unified fine-grained time step is used, although the scheduling accuracy can be improved, it will cause a "variable explosion" in the number of nodes and arcs in the spatio-temporal network, resulting in too long algorithm solving time in large-scale task scheduling scenarios and difficult to meet the actual needs of airport real-time scheduling.
[0004] In-depth analysis shows that the core problem of the existing technology being unable to balance the above requirements lies in the failure to establish an association mechanism between time step adjustment and task priorities, and unable to achieve differential processing of "accurate description of key tasks and streamlined calculation of non-critical periods". This defect directly leads to the difficulty of the existing solutions to balance the accuracy guarantee of high-priority tasks and the computational efficiency of large-scale scenarios. Therefore, how to reduce the computational redundancy of non-critical periods by adaptively adjusting the time granularity while ensuring the accurate satisfaction of the time constraints of high-priority tasks has become the core technical challenge urgently to be solved in the optimization of the current fresh water truck scheduling algorithm. Summary of the Invention
[0005] The purpose of the present invention is to provide a scheduling method combining event priorities, which has the advantages of significantly improving the overall solving efficiency of the algorithm while ensuring the scheduling accuracy of high-priority tasks, and adapting to the needs of large-scale real-time scheduling.
[0006] To achieve the above objectives, this invention provides a scheduling method that combines event priority, overcoming the deficiency in existing technologies where "accuracy and efficiency cannot be simultaneously achieved" when solving the water truck scheduling problem using spatiotemporal networks. The method includes: Step S1, initial parameter configuration and sparse network construction to obtain an initial scheduling scheme. Step S2, for the initial scheduling scheme Step S3: Perform priority-driven time conflict detection and record conflict nodes; Step S4: Insert conflict nodes and reconstruct the network to obtain a new scheduling scheme; Step S5: Adjust the time step based on the new scheduling scheme; Step S6: Iteratively converge to the adjusted scheduling scheme. If the convergence is satisfied, output the result; otherwise, return to Step S4.
[0007] Preferably, step S1 includes: step S11, inputting the full set of scheduled tasks. , Where i and N represent indicator algebras, which have no actual computational meaning;
[0008] Step S12, obtain them one by one The core parameters include: task priority Time window parameters, maximum allowed dwell time Maximum number of iterations , threshold for improvement of objective function Step S13: Filter the highest priority tasks ( The key time points form the initial time node set. Among them, the initial time node set Includes: the earliest possible start time of the task The latest time the task needs to be completed Initial departure time of the water truck Stop time at water replenishment stations Step S14, based on The node layer of the spatiotemporal network is constructed, and the arc layer is constructed based on the constraints of the water truck's travel route and the operation sequence, forming the initial sparse spatiotemporal network. The operation sequence constraint constructs an arc layer including: a travel arc, an operation arc, and a water replenishment arc; step S15, for the initial sparse spatiotemporal network Construct a relaxation optimization model and solve for the initial scheduling scheme. The initial scheduling scheme includes multiple schemes, each of which includes the vehicle travel route, the planned completion time of each task, and the planned dwell time.
[0009] Preferably, the relaxation optimization model includes: retained core constraints, such as the time window constraints of high-priority tasks and the water replenishment duration constraints of the water truck, to ensure that the core requirements do not deviate; relaxed non-core constraints, such as temporarily relaxing the continuity constraints of vehicle travel paths and the rigid constraints of dwell time for low-priority tasks, without pursuing the optimal solution, only needing to obtain an initial solution "that can be used for subsequent verification"; and a solution logic that, guided by the comprehensive optimization objective of the scheduling scheme, uses conventional optimization algorithms to solve the relaxation model, ultimately outputting the initial scheduling scheme. .
[0010] Preferably, step S2 includes: step S21, setting the initial scheduling scheme. All the solutions are ordered from highest to lowest task priority, where, , For the highest priority level, time conflict detection is performed one by one; in step S22, the core information of all conflict nodes is recorded, including conflict type, conflict occurrence time, and the priority of the corresponding task. .
[0011] Preferably, in step S21, the time conflict detection includes: a time conflict is determined if any of the following dimensions are met:
[0012] Dimension 1: Calculating the actual arrival time of the vehicle at the task Time of execution With the planned arrival time deviation ,like , Priority If the corresponding baseline time step is used, it is determined to be a time conflict;
[0013] Dimension 2: Calculating vehicle dwell time ,in, For the task The actual start time, if , Priority If the maximum allowed dwell time for a task is not specified, it is considered a time conflict.
[0014] Preferably, step S3 includes: step S31, forming a new node set from all conflicting nodes. Step S32, add the new node set By incorporating the time node set from the previous iteration, we obtain the updated time node set. ;in, For the first The set of time nodes for the next iteration. For the first The set of time nodes for the next iteration. The current iteration number; Step S33, based on the updated set of time nodes. Reconstruct the arc layer of the spatiotemporal network to obtain the updated spatiotemporal network. ;in, Representing the The spatiotemporal network of the next iteration; step S34, based on The optimization model is solved again to obtain a new scheduling scheme. ,in, Representing the The scheduling scheme for the next iteration.
[0015] Preferably, step S4 includes: step S41, applying a new scheduling scheme. Perform dense segment encryption and sparse segment thinning: Step S42, for the transition period between dense and sparse segments, use linear interpolation to adjust the time step size to avoid solution fluctuations caused by sudden changes in step size and ensure the stability of the algorithm solution.
[0016] Preferably, dense segment encryption includes: if within a certain time period, a unit time contains segments of the same priority... Number of tasks ;in, The density threshold is used; the time step of this period is compressed to... ;in, Priority The corresponding minimum time step; sparse segment thinning includes: if there are no tasks executed within a certain time period, or only long-distance travel tasks, then the time step of that time period is adjusted to Δt. max =30min; where Δt max This represents the maximum time step.
[0017] Preferably, step S5 includes: satisfying any of the following convergence conditions; if satisfied, output the result; otherwise, return to step S4; Convergence condition 1: Current scheduling scheme There are no time conflicts, meaning all tasks meet the requirements. and Convergence condition 2: Number of iterations Reaching the preset maximum number of iterations Convergence condition 3: The difference between the objective functions of two adjacent iterations ;in, For the first The objective function value of the next iteration. For the first The objective function value of the next iteration. The preset improvement threshold.
[0018] In summary, compared with the prior art, the scheduling method combining event priority provided by the present invention has the following beneficial effects:
[0019] First, this invention achieves differentiated characterization of the time dimension through an adaptive time step adjustment strategy of "priority layering + density adaptation"—a fine-grained step of 1 to 2 minutes is used to ensure accuracy in high-priority task-intensive segments, while a coarse-grained step of 30 minutes is used to simplify calculations in blank or sparsely driven segments. This effectively solves the defects of "insufficient accuracy" or "low efficiency" of the traditional uniform step size across the entire domain, and significantly improves the overall performance of the algorithm.
[0020] Second, this invention ensures that the time window constraints of high-priority tasks are met with the highest priority through the logical design of "detecting conflicts from high to low priority" and "locking high-priority tasks with the minimum step size". This reduces the risk of default for high-priority tasks and is suitable for the actual needs of "prioritizing key flights" in airport water truck scheduling.
[0021] Third, this invention, through its "on-demand encryption" mechanism, increases the number of nodes only in conflict areas and high-priority task areas, avoiding the "variable explosion" caused by global fine-grained discretization. Compared to traditional global fine-grained discretization methods, this invention can reduce the number of decision variables and state variables by 30% to 60%, enabling the rapid generation of large-scale real-time scheduling schemes.
[0022] Fourth, the priority hierarchical mechanism and step size adjustment threshold of the present invention can be flexibly configured according to the actual scheduling scenario. It can adapt to small-scale precise scheduling needs as well as large-scale real-time scheduling needs, and can still maintain stable solution performance in scenarios with fluctuating task numbers and dynamic priority adjustments. Attached Figure Description
[0023] Figure 1 This is a flowchart of a scheduling method combining event priorities proposed in this invention. Detailed Implementation
[0024] The following will be combined with the appendix in the embodiments of the present invention. Figure 1 The technical solutions, structural features, objectives and effects achieved in the embodiments of the present invention will be described in detail.
[0025] It should be noted that the accompanying drawings are in a very simplified form and use non-precise proportions. They are only used to facilitate and clarify the purpose of illustrating the embodiments of the present invention, and are not intended to limit the implementation conditions of the present invention. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportional relationship, or adjustments to the size should still fall within the scope of the technical content disclosed in the present invention, provided that they do not affect the effects and objectives that the present invention can produce.
[0026] It should be noted that, in this invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only the expressly listed elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.
[0027] like Figure 1 As shown, this invention proposes a scheduling method that combines event priorities, aiming to overcome the defect in the prior art that "accuracy and efficiency cannot be achieved simultaneously" when solving the water truck scheduling problem using spatiotemporal networks;
[0028] The method includes:
[0029] Step S1: Initial parameter configuration and sparse network construction to obtain the initial scheduling scheme. ;
[0030] Step S2, for the initial scheduling scheme Perform priority-driven time conflict detection and record conflict nodes;
[0031] Step S3: Insert conflicting nodes and reconstruct the network to obtain a new scheduling scheme;
[0032] Step S4: Adjust the time step based on the new scheduling scheme;
[0033] Step S5: Iteratively determine the convergence of the adjusted scheduling scheme. If the condition is met, output the result; otherwise, return to step S4.
[0034] Specifically, step S1 includes:
[0035] Step S11, input the full set of scheduled tasks. , Where i and N represent indicator algebras, which have no actual computational meaning;
[0036] Step S12, obtain them one by one The core parameters include: task priority Time window parameters (earliest start time EST, latest completion time LFT), maximum allowed dwell time (Set in layers according to priority, where ω is the task priority), maximum number of iterations , threshold for improvement of objective function ;
[0037] Step S13, filter the highest priority tasks ( The key time points form the initial time node set. Among them, the initial time node set Includes: the earliest possible start time of the task The latest time the task needs to be completed Initial departure time of the water truck Stop time at water replenishment stations ;
[0038] Step S14, based on The node layer of the spatiotemporal network is constructed, and the arc layer is constructed based on the constraints of the water truck's travel route and the operation sequence, forming the initial sparse spatiotemporal network. The task sequence constraint construction arc layer includes: a travel arc (connecting the water truck travel status nodes corresponding to adjacent time nodes, updating the travel time, distance, and other attributes of the arc segment) and a task arc (connecting the time nodes corresponding to task execution, associating tasks). of , (Equal constraint parameters), water replenishment arc (connecting the time node corresponding to the water replenishment station of the water truck, and associating the water replenishment duration constraint parameters).
[0039] Step S15, for the initial sparse spatiotemporal network Construct a relaxation optimization model and solve for the initial scheduling scheme. The initial scheduling scheme includes multiple schemes, each of which includes the vehicle travel route, the planned completion time of each task, and the planned dwell time.
[0040] The core design of the relaxation optimization model is to "preserve core constraints and relax non-core constraints". The purpose is to reduce the initial solution complexity, quickly obtain feasible scheduling schemes, and provide a foundation for subsequent conflict detection and iterative optimization.
[0041] The specific explanation is as follows: Retained core constraints: High-priority tasks ( Time window constraint () ), water truck refill time constraints, ensuring core needs remain unchanged; relaxed non-core constraints, temporarily relaxing vehicle travel path continuity constraints and low-priority tasks ( The rigid constraint on the dwell time of the scheduling scheme does not require pursuing the optimal solution; only an initial scheme that can be used for subsequent verification is needed. The solution logic is based on the comprehensive optimization objective of the scheduling scheme. Guided by the principle of minimizing the total mileage, a conventional optimization algorithm is used to solve the relaxation model, ultimately outputting an initial scheduling scheme. .
[0042] Specifically, step S2 includes:
[0043] Step S21, for the initial scheduling scheme All solutions are ordered from highest to lowest task priority. , (Assuming the highest priority), time conflict detection is performed one by one;
[0044] Step S22: Record the core information of all conflict nodes. The core information includes the conflict type (arrival time deviation conflict / staying timeout conflict) and the time when the conflict occurred (i.e., the actual arrival time). Or the actual start time ), priority of corresponding tasks .
[0045] In step S21, the time conflict detection includes determining a time conflict if any of the following dimensions are met:
[0046] Dimension 1: Calculating the actual arrival time of the vehicle at the task Time of execution With the planned arrival time deviation ,like , Priority If the corresponding baseline time step is not specified, then a time conflict is determined.
[0047] Dimension 2: Calculating vehicle dwell time ,in, For the task The actual start time, i.e., the time when the vehicle is on the mission. The start time of the operation at the execution location, if , Priority If the maximum allowed dwell time for a task is not specified, it is considered a time conflict.
[0048] Specifically, step S3 includes:
[0049] Step S31: Combine all conflicting nodes into a new node set. ;
[0050] Step S32, add the new node set By incorporating the time node set from the previous iteration, we obtain the updated time node set. ;in, For the first The set of time nodes for the next iteration. For the first The set of time nodes for the next iteration. This represents the current iteration number;
[0051] Step S33, based on the updated set of time nodes Reconstruct the arc layer of the spatiotemporal network to obtain the updated spatiotemporal network. ;in, Representing the The spatiotemporal network of the next iteration;
[0052] As previously mentioned, the updated spatiotemporal network This includes: the driving arc, the working arc, and the water replenishment arc;
[0053] Step S34, based on The optimization model is solved again to obtain a new scheduling scheme. ,in, Representing the The scheduling scheme for the next iteration.
[0054] Specifically, step S4 includes:
[0055] Step S41, for the new scheduling scheme Perform dense segment encryption and sparse segment thinning:
[0056] Specifically, dense segment encryption includes: if within a certain time period, there are multiple segments of the same priority within a unit of time... Number of tasks ;in, This is a density threshold used to determine whether the task distribution is dense, or whether the time period contains high-priority tasks. For tasks, the time step of that time period is compressed to... ;in, Priority The corresponding minimum time step is the minimum step size threshold that guarantees the accuracy of this priority task.
[0057] in time Δt max=12min , hour As priority increases; such as The time is 35 minutes, ensuring accurate depiction of the task time constraints within this period;
[0058] Specifically, sparse segment thinning includes: if there are no tasks executed within a certain time period, or only long-distance travel tasks, then the time step of that time period is adjusted to Δt. max =30min; where Δt max The maximum time step is the threshold for controlling computational redundancy, which reduces the number of nodes and arcs within that time period and reduces computational redundancy.
[0059] Step S42: For the transition period between dense and sparse segments, linear interpolation is used to adjust the time step to avoid solution fluctuations caused by sudden changes in the step size and to ensure the stability of the algorithm solution.
[0060] Specifically, step S5 includes: if any of the following convergence conditions are met, output the result; otherwise, return to step S4.
[0061] Convergence condition 1: Current scheduling scheme ( There are no time conflicts in the number of iterations, meaning all tasks satisfy the condition. and ;
[0062] Convergence condition 2: Number of iterations Reaching the preset maximum number of iterations ;
[0063] Convergence condition 3: The difference in the objective function between two adjacent iterations ;in, For the first The objective function value of the next iteration. For the first The objective function value of the next iteration. The preset improvement threshold is set so that the algorithm tends to stabilize when the objective function does not improve significantly.
[0064] Although the present invention has been described in detail through the preferred embodiments above, it should be understood that the above description should not be considered as a limitation of the present invention. Various modifications and substitutions to the present invention will be apparent to those skilled in the art after reading the above description. Therefore, the scope of protection of the present invention should be defined by the appended claims.
Claims
1. A scheduling method combining event priority, characterized in that, The method overcomes the limitation of existing technologies in solving the water truck scheduling problem using spatiotemporal networks, where "accuracy and efficiency cannot be simultaneously achieved"; the method includes: Step S1: Initial parameter configuration and sparse network construction to obtain the initial scheduling scheme. ; Step S2, for the initial scheduling scheme Perform priority-driven time conflict detection and record conflict nodes; Step S3: Insert conflicting nodes and reconstruct the network to obtain a new scheduling scheme; Step S4: Adjust the time step based on the new scheduling scheme; Step S5: Iteratively determine the convergence of the adjusted scheduling scheme. If the condition is met, output the result; otherwise, return to step S4.
2. The scheduling method combining event priorities according to claim 1, characterized in that, Step S1 includes: Step S11, input the full set of scheduled tasks. , Where i and N represent indicator algebras, which have no actual computational meaning; Step S12, obtain them one by one The core parameters include: task priority Time window parameters, maximum allowed dwell time Maximum number of iterations , threshold for improvement of objective function ; Step S13, filter the highest priority tasks ( The key time points form the initial time node set. Among them, the initial time node set Includes: the earliest possible start time of the task The latest time the task needs to be completed Initial departure time of the water truck Stop time at water replenishment stations ; Step S14, based on The node layer of the spatiotemporal network is constructed, and the arc layer is constructed based on the constraints of the water truck's travel route and the operation sequence, forming the initial sparse spatiotemporal network. The operation sequence constraint constructs an arc layer including: travel arc, operation arc, and water replenishment arc. Step S15, for the initial sparse spatiotemporal network Construct a relaxation optimization model and solve for the initial scheduling scheme. The initial scheduling scheme includes multiple schemes, each of which includes the vehicle travel route, the planned completion time of each task, and the planned dwell time.
3. The scheduling method combining event priorities according to claim 2, characterized in that, The relaxation optimization model includes: retained core constraints, such as the time window constraints for high-priority tasks and the water truck refilling time constraints, to ensure that core requirements do not deviate; relaxed non-core constraints, such as temporarily relaxing the continuity constraints of vehicle travel paths and the rigid constraints of dwell time for low-priority tasks, without pursuing the optimal solution, only needing to obtain an initial solution "usable for subsequent verification"; and a solution logic that, guided by the comprehensive optimization objective of the scheduling scheme, uses conventional optimization algorithms to solve the relaxation model, ultimately outputting the initial scheduling scheme. .
4. The scheduling method combining event priorities according to claim 3, characterized in that, Step S2 includes: Step S21, for the initial scheduling scheme All the solutions are ordered from highest to lowest task priority, where, , Assuming the highest priority, time conflict detection is performed one by one; Step S22: Record the core information of all conflict nodes. The core information includes the conflict type, the time when the conflict occurred, and the priority of the corresponding task. .
5. A scheduling method combining event priorities according to claim 4, characterized in that, In step S21, the time conflict detection includes: a time conflict is determined if any of the following dimensions are met: Dimension 1: Calculating the actual arrival time of the vehicle at the task Time of execution With the planned arrival time deviation ,like , Priority If the corresponding baseline time step is used, it is determined to be a time conflict; Dimension 2: Calculating vehicle dwell time ,in, For the task The actual start time, if , Priority If the maximum allowed dwell time for a task is not specified, it is considered a time conflict.
6. A scheduling method combining event priorities according to claim 5, characterized in that, Step S3 includes: Step S31: Combine all conflicting nodes into a new node set. ; Step S32, add the new node set By incorporating the time node set from the previous iteration, we obtain the updated time node set. ;in, For the first The set of time nodes for the next iteration. For the first The set of time nodes for the next iteration. This represents the current iteration number; Step S33, based on the updated set of time nodes Reconstruct the arc layer of the spatiotemporal network to obtain the updated spatiotemporal network. ;in, Representing the The spatiotemporal network of the next iteration; Step S34, based on The optimization model is solved again to obtain a new scheduling scheme. ,in, Representing the The scheduling scheme for the next iteration.
7. A scheduling method combining event priorities according to claim 6, characterized in that, Step S4 includes: Step S41, for the new scheduling scheme Perform dense segment encryption and sparse segment thinning: Step S42: For the transition period between dense and sparse segments, linear interpolation is used to adjust the time step to avoid solution fluctuations caused by sudden changes in the step size and to ensure the stability of the algorithm solution.
8. A scheduling method combining event priorities according to claim 7, characterized in that, Dense segment encryption includes: if within a certain time period, there are multiple segments of the same priority within a unit of time. Number of tasks ;in, The density threshold is used; the time step of this period is compressed to... ;in, Priority The corresponding minimum time step; Sparse segment thinning includes: if there are no tasks executed within a certain time period, or only long-distance travel tasks, then the time step of that time period is adjusted to Δt. max =30min; where Δt max This represents the maximum time step.
9. A scheduling method combining event priorities according to claim 8, characterized in that, Step S5 includes: satisfying any of the following convergence conditions; if satisfied, output the result; otherwise, return to step S4. Convergence condition 1: Current scheduling scheme There are no time conflicts, meaning all tasks meet the requirements. and ; Convergence condition 2: Number of iterations Reaching the preset maximum number of iterations ; Convergence condition 3: The difference in the objective function between two adjacent iterations ;in, For the first The objective function value of the next iteration. For the first The objective function value of the next iteration. The preset improvement threshold.