A flight departure guarantee resource configuration scheduling optimization management method

Abstract

The present application relates to the technical field of flight management, in particular to a kind of flight departure guarantee resource configuration scheduling optimization management method, comprising: according to the runway, run sequence and time interval corresponding to flight schedule list, flight schedule list is constructed in the form of front order and rear order with flight state diagram;With the maximum capacity and minimum adjustment time interval of each node in flight state diagram in unit time period as constraint condition, determine the transition cost of each state in flight state transition process, sort each flight into time-adjusted flight list with transition cost;Iterative solution is carried out using time-adjusted flight list, determine the optimal adjustment time interval in unit time, determine the adjustment quantity in unit time period according to adjustment time interval combined with time slot allocation;When adjustment quantity meets preset constraint rule, corresponding flight is combined with data, and the combined data is regarded as the output adjustment scheme;The efficiency and executability of flight scheduling processing are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of flight management, in particular to a flight departure guarantee resource configuration scheduling optimization management method. BACKGROUND

[0002] In airport flight scheduling work, due to large flight volume and complex running influencing factors, the departure time of the flight is usually planned in advance. In the case of adverse weather, emergency rescue and other situations, the executable upper limit of the whole field flight in the affected period is determined, and according to the time adjustment application of the airline, the first-come-first-served principle is approved when the adjustment time reaches the upper limit of the period. When the adjustment time reaches the upper limit of the period, stop approving; such processing is easy to make multiple flight scheduling difficult, and reduce the overall operation order maintenance and operation efficiency of the airport in the case of large-area flight delay.

[0003] For example, Chinese patent publication No. CN115983613A discloses a flight splicing method, device, electronic equipment and storage medium. The flight splicing method comprises: if the type of the current business stage is a flight planning stage, the planned flights that have completed flight splicing in the current flight date are disconnected and spliced, and flight splicing is completed based on the aircraft number information, planned arrival time and planned departure time of each planned flight; if the type of the current business stage is a flight dynamic stage, a plurality of dynamic flights that have not been cancelled and have not ended are determined, the dynamic flights that have completed flight splicing in the plurality of dynamic flights are disconnected and processed, and flight splicing is completed based on the aircraft number information, priority order table of arrival time and priority order table of departure time of each dynamic flight.

[0004] For example, Chinese patent publication No. CN118798565A discloses a decision method, device and equipment for sorting departing flights, and storage medium. By acquiring the flight state and flight information, the first state duration and the first departure strategy of the to-be-planned flight entering the next state are determined. Then, the pre-activated flights, the flights in the to-be-departed state and the flights in the departing state are constructed into pre-activated queues, to-be-departed queues and departing queues respectively, and sorted according to the first state duration to obtain first waiting queue information. Subsequently, the first waiting queue information and the corresponding departure strategy are input into the sorting optimization model to obtain the decision optimization suggestion for the departure flight sorting.

[0005] The prior art obtains the output flight list by splicing the arrival time and departure time of the flight across the night, and sorts the flight queue by the flight pre-activation state, departure state and other states to obtain the departure flight decision data; these methods represent the processing method of the flight in the sorting process, but the flight needs to consider the flight list to be dispatched when sorting, determine the configuration of the departure guarantee resource of the flight under multiple constraint conditions, and finally realize the global optimality of the flight dispatching and the executability of the scheme. SUMMARY

[0006] To solve the above technical problems, the technical scheme adopted by the present application is: a flight departure guarantee resource configuration scheduling optimization management method, comprising: S1, data extraction is performed according to the current departure and arrival flights to obtain a flight schedule list.

[0007] S2, according to the runway, run sequence and time interval corresponding to the flight schedule list, a flight state diagram is constructed in a pre-sequence and post-sequence manner.

[0008] S3, when the state of the flight state diagram changes, the maximum capacity and minimum adjustment time interval of each node in the flight state diagram within a unit time period are taken as constraint conditions, the transition cost of each state in the flight state diagram is determined, and each flight is sorted into a dispatching flight list according to the transition cost.

[0009] S4, the dispatching flight list is used for iterative solution to determine the optimal adjustment time interval within a unit time, and the adjustment quantity within a unit time period is determined according to the adjustment time interval combined with the time slot allocation.

[0010] S5, when the adjustment quantity meets the preset constraint rule, the corresponding flight is combined, and the combined data is regarded as an output adjustment scheme.

[0011] The present application has the following advantages: first, the present application combines the extracted arrival schedule and departure schedule into a flight state diagram by scheduling the associated data of the pre-sequence flight and the post-sequence flight, clearly defines the correlation between different flights and the current flight, prevents the problem of pre-sequence and post-sequence conflict during flight scheduling, and provides a data basis for subsequent flight calculation.

[0012] Second, the present application takes the maximum capacity and minimum adjustment time interval of a unit time period as a constraint, constructs a penalty factor combined with the dispatching range, dispatching ratio and delay cost, generates a dispatching flight list through the transition cost, preliminarily concentrates the multiple constraint problems into the constraints pointed by the dispatching range, dispatching ratio and delay cost, determines the cost required to record each flight scheduling, and improves the feasibility of the preliminary scheduling scheme.

[0013] Thirdly, the optimal solution of the local sub-constraint is mapped to multiple time periods by combining the piece identification conflict processing, and finally the adjustment time interval and the adjustment quantity under the compliance of multiple constraints are determined, so as to avoid the chain reaction of new delay caused by time adjustment, and finally the adjustment quantity is synchronized to the flight status diagram, and the preset constraint rule is used for re-comparison, so as to ensure the consistency of the scheme before and after. BRIEF DESCRIPTION OF DRAWINGS

[0014] The application will be further described below in combination with the drawings and embodiments.

[0015] Figure 1 It is a process schematic diagram of a flight departure guarantee resource configuration scheduling optimization management method.

[0016] Figure 2 It is a process schematic diagram of step S1 of a flight departure guarantee resource configuration scheduling optimization management method.

[0017] Figure 3 It is a process schematic diagram of step S2 of a flight departure guarantee resource configuration scheduling optimization management method.

[0018] Figure 4 It is a process schematic diagram of step S3 of a flight departure guarantee resource configuration scheduling optimization management method.

[0019] Figure 5 It is a process schematic diagram of step S4 of a flight departure guarantee resource configuration scheduling optimization management method. DETAILED DESCRIPTION

[0020] The embodiments of the application will be described in detail below. The embodiments described below are exemplary and are only used to explain the application, and cannot be understood as a limitation of the application. If the specific technology or condition is not indicated in the embodiments, the technology or condition described in the literature in the art or according to the product instruction is used.

[0021] Reference Figure 1 A flight departure guarantee resource configuration scheduling optimization management method comprises the following steps: S1, data extraction is performed according to the current departure and arrival flights to obtain a flight schedule list.

[0022] S2, a flight status diagram is constructed in a pre-sequence and post-sequence manner according to the runway, runway sequence and time interval corresponding to the flight schedule list.

[0023] S3, when the state of the flight status diagram changes, the maximum capacity and the minimum adjustment time interval of each node in the flight status diagram within a unit time period are taken as constraint conditions, the transfer cost of each state in the flight status diagram is determined, and each flight is sorted into a time adjustment flight list according to the transfer cost.

[0024] S4, using the timed flight list to iteratively solve, determine the optimal adjustment time interval per unit time, according to the adjustment time interval combined with time slot allocation, determine the adjustment number per unit time period.

[0025] S5, when the adjustment number meets the preset constraint rule, the corresponding flight is combined, and the combined data is regarded as the output adjustment scheme.

[0026] The current scheme describes the pre-sequencing-posteriori association by constructing a flight state diagram, sorts the timed flights according to the transition cost of each state, iteratively solves the optimal adjustment time interval, and essentially is the optimization of timed decision-making under the dynamic change of ground state.

[0027] When extracting the data of the flight, the corresponding data will be spliced up and down from the arrival and departure flight data, and the flights of the current scheduling configuration are determined to belong to the interval analyzed.

[0028] As shown in Figure 2 The implementation of step S1 includes: S11, obtaining the number of flights to be scheduled at any time, when the number of flights to be scheduled meets the scheduling-out ratio, splicing the arrival and departure flight data up and down, and recording the data start and end time interval at the time of splicing; The scheduling-out ratio refers to the ratio of the number of normally scheduled flights and the number of abnormally scheduled flights to the total number of flights, that is, the ratio of the number of normally scheduled flights and the number of abnormally canceled flights to the total number of flights; The current scheme tends to schedule flights in peak hours and bad weather in resource-intensive or safety-sensitive scenarios, and the number of scheduled flights needs to be less than the scheduling-out ratio to prevent over-scheduling of resources, and the scheduling-out ratio is 0.1 in general scenarios.

[0029] S12, according to the planned takeoff time and the actual takeoff time in the data start and end time interval, the spliced flight data is removed, and the flight schedule list is output.

[0030] According to the planned takeoff time and the actual takeoff time, some flights are removed, including: departure flights with actual takeoff time and takeoff time not later than the engine trigger time, and STD plan to remove the wheel file time in the data start and end time interval.

[0031] Arrival flights with actual landing time and landing time not later than the engine trigger time, and STA plan to land in the data start and end time interval.

[0032] The engine trigger time generally refers to a reference time point of starting an airplane engine or starting a key operation in the field of aviation. The planned wheel chock removal time STD is a time point at which the airplane is pushed out from a parking position, the wheel chock is removed, and the airplane starts to taxi. The data start and end time interval refers to a continuous time period containing key time nodes (such as STD and STA) of a flight plan and actual running time (such as actual takeoff / landing time). The length of the time period is not less than 6 hours, and the time period is used for prior scheduling of the flight.

[0033] In an embodiment of the present application, the flight status graph classifies each flight according to a corresponding runway and run sequence, records special weather and airspace activities related to the current flight, and records a corresponding timestamp.

[0034] In step S2, when the flight status graph is constructed, the nodes not only contain flight statuses, but also are associated with occupied ground station resources and parking position resources, and emphasize resource scheduling in the ground station process to realize scheduling of corresponding flight. The weight of the edge not only considers a run sequence interval, but also contains ground station guarantee time conflicts and parking position reuse conflicts, so that a previous flight and a subsequent flight meet constraints of the ground station process. For example, if a previous flight occupies P1 to 10:00 and a subsequent flight needs to use P1, the subsequent flight must be greater than or equal to 10:00+30 minutes of cleaning time.

[0035] As shown in FIG. 2, the implementation of step S2 includes the following steps. Figure 3 S21, it is judged whether a previous flight exists for a current flight. If the previous flight exists, a previous flight and a subsequent flight of the current flight are modeled, and a scheduling state of each flight is determined according to data information of the runway, the run sequence, and the time interval of the previous flight and the subsequent flight. The scheduling state directly contains resource occupation of the runway and the parking position of the current flight, and a run sequence position and a takeoff interval of each flight.

[0036] S22, if the previous flight does not exist, a scheduling state of each flight is determined according to the runway, the run sequence, and the time interval of the subsequent flight of the current flight and other flights. The absence of the previous flight represents that the current flight is the first takeoff in a day, and how the flight departs is recorded according to content configured by the subsequent flight.

[0037] S23, a shortest time interval between scheduling states is regarded as a target cycle time, and a scheduling mode of the target cycle time is regarded as a state change of the flight status graph.

[0038] The previous flight refers to a last flight segment performed by the same airplane before the current flight is performed. For example, if a flight plan of an airplane is “Guangzhou-Chengdu-Guangzhou”, “Guangzhou-Chengdu” is a previous flight of “Chengdu-Guangzhou”. The state of the previous flight directly affects execution of the subsequent flight.

[0039] The subsequent flight refers to the next flight segment performed by the same aircraft after the current flight task. For example, "Chengdu to Guangzhou" is the subsequent flight of "Guangzhou to Chengdu". The execution of the subsequent flight depends on the resource release of the previous flight and the timely connection of transfer passengers.

[0040] The scenario of the current flight dispatch is used to identify the multi-segment ticket usage rules in the aviation industry, that is, not taking the previous flight will cause the subsequent flight to be cancelled, and the on-time rate of the current flight is determined by querying the status of the previous flight of the same aircraft, and the status of the current flight to be dispatched is formed according to the on-time rate. The flight status diagram is combined into multiple nodes.

[0041] As for the dispatch mode, it represents the data of how the related flight is dispatched within the corresponding time interval, such as delaying for a few minutes, adjusting the take-off time for a few minutes, and these data are recorded in the form of current flight status change.

[0042] The shortest time interval between the dispatch states refers to the time difference between the end state of the previous flight and the start state of the current flight, and the end state of the current flight and the start state of the subsequent flight, which reflects the minimum interval of airport ground dispatch and resource recovery time. This value will focus on the process of resource reuse and release, and explain the time interval between the current flight and the next task after completing the task.

[0043] Preferably, the implementation of step S23 further comprises: generating a transfer time matrix composed of different previous flights and current flights according to the connection relationship between the previous flights and the current flight; and recording the minimum transfer time of different previous flights and current flights in the transfer time matrix.

[0044] The transfer time matrix is associated with the flight schedule list corresponding to the current flight, and the expected take-off time of the outbound flight is updated.

[0045] After updating the expected take-off time of the outbound flight, the updated time is also synchronized to the subsequent flight to determine whether the subsequent flight can pass the latest state of the current flight.

[0046] The purpose of connecting the previous and subsequent flights through the flight status diagram is to check the influence of one flight and its associated other flights such as runway flights, transfer flights, and corresponding flights on the configuration scenario of the previous and subsequent flights by associating the actual execution time of a single flight with the time adjustment range of the original planned time. These previous and subsequent flights are identified and sorted by time period to measure the influence of a single flight on the overall flight time coordination when scheduling.

[0047] In one embodiment of the present application, in the transition cost calculation of each state, real-time deviation is introduced; for example, if the actual landing time of the previous flight is 10 minutes later than the plan, the transition cost of the subsequent flight needs to be added, and when the actual state deviates from the plan by more than a threshold value, the transition cost is automatically recalculated and the scheduled flight list is updated.

[0048] Meanwhile, the maximum capacity represented in step S3 is not for a single flight, but refers to the attribute of the resource (runway / stand) occupied by the flight, at which time the flight and the maximum capacity of the corresponding runway or stand in a unit time period are checked to determine the current flight configuration.

[0049] As shown in Figure 4 The implementation of step S3 includes: S31, in response to changes in the current flight state diagram, a state change signal is generated, and the state change signal is constructed according to the connection relationship of the current flight state diagram.

[0050] S32, each flight under state transition is read by logical operation, to determine whether there is a conflict between the previous and subsequent flights and the current flight, and the data after the determination is used to construct a state transition matrix.

[0051] S33, the scheduling range, the proportion of scheduling out and the delay cost at each scheduling are used as penalty factors to construct the transition cost of the state transition matrix in multiple time periods.

[0052] S34, based on the transition cost of the state transition matrix, the transition probability that meets the constraints in each time period is calculated, and the transition probability value is output from large to small as a scheduled flight list.

[0053] When the flight state changes, including temporary replacement of the current flight model, sudden increase in passenger transfer volume, previous delay, temporary runway closure and stand personnel shortage, etc., the state of the flight record may also change due to different in-building transfers and inter-building transfers, and different transit times need to be set. At this time, the part of the flight state change is recorded, which is marked with a state change signal, and the conflict risk of the current flight is determined according to the resources and time of the previous and subsequent flights of the current flight, and the adjustment cost is calculated to select the transition probability that meets the constraints, and the flight list is output according to the probability value. The essence is to capture the uncertainty of the allocation of the previous and subsequent flights, to avoid the problem of unknown adjustment cost caused by subjective judgment during flight scheduling.

[0054] When the node or edge of the flight status graph changes, automatically generate a standardized signal, the signal needs to include three elements of change source, change type, change amplitude; For example, the change source is the number of the previous flight, the change type records the time change form of the delay, and the change amplitude records the relative time of the delay; A complete state change signal can be represented as, such as flight F1 number CA1234, Beijing to Shanghai, the actual landing time is later than the plan, delayed for 10 minutes; If the current flight situation is explained, the record will record the relative delay time of take-off, and the runway may be adjusted and the expected take-off time.

[0055] When constructing state transitions according to connection relationships, the state transition is a chain of associated chains such as previous sequence→current sequence→next sequence, and the change of each flight is transmitted to the associated flight to obtain a transition path.

[0056] In the logical operation reading in step S32, the logical operation is based on resource constraints (runway / stand capacity), time constraints (minimum interval / overhaul time), to judge whether the flight state transition is compliant, and record the conflict and non-conflict parts of each flight.

[0057] If the number of flights in a unit period that the flight intends to occupy the resource (such as runway 1) is greater than the maximum capacity, record the conflict and mark 1; Otherwise, it is compliant and marked 0; If the time difference between the previous and next sequence flights is less than the overhaul time, record the conflict and mark 1; Otherwise, it is compliant and marked 0.

[0058] When constructing the state transition matrix, the matrix is a two-dimensional table of current state×next state, the row represents the current state of the flight, such as waiting for take-off, and the column represents the next state to be transferred, such as taking off after adjusting the time by 5 minutes, while marking the conflict type.

[0059] The time adjustment amplitude records the specific value of the flight time adjustment, such as delaying 2 minutes, advancing 2 minutes, etc.

[0060] When constructing the penalty factor, the current required time adjustment amplitude, time adjustment ratio, and delay cost are used as the main part of the judgment, and a multi-time window form is used to process the flights within every 15 minutes, and the transition cost is recorded respectively, and then according to the transition cost in each time window, an exponential value is converted, and the ratio of the exponential value to the total period is used as the transition probability.

[0061] Preferably, the implementation of step S33 further includes: the time adjustment amplitude points to the time adjustment penalty cost and the departure balance difference cost, the time adjustment ratio points to the ground departure difference penalty cost, and the delay cost points to the flight delay penalty cost, the flight cancellation penalty cost and the active delay penalty cost. The sum of the penalty costs of each time period is used as the transition cost of the state transition matrix. The multiple costs directly pointed to by the three penalty factors will represent the current main optimization part.

[0062] The time adjustment penalty cost is mainly set according to the aircraft type of each flight and the priority of the airline, and a time adjustment penalty coefficient is set. The product of the time adjustment penalty coefficient and the absolute value of the time adjustment range is set as the time adjustment penalty cost of the corresponding flight. The time adjustment penalty coefficient can be directly obtained from historical data, which mainly reflects the positive correlation between the current time adjustment penalty cost and the time adjustment size.

[0063] The ground agent dispatch difference penalty cost considers the deviation of the actual dispatch ratio from the target ratio. At this time, the dispatch ratio records the number of canceled and dispatched flights, and verifies the real-time dispatch ratio of each ground agent. The actual dispatch ratio obtained in each time period is used as the actual dispatch ratio at this time. The penalty cost of each flight belonging to the ground agent is explained. The weight of the dispatch penalty cost is set according to the ratio of the number of flights covered by each ground agent to the total number of flights in the corresponding time period. The square value of the deviation of the actual dispatch ratio from the target ratio is multiplied by the weight to obtain the ground agent dispatch difference penalty cost. This ground agent dispatch difference penalty cost is normalized according to all values, which indicates the overall deviation of the flight belonging to the ground agent. The target ratio is set to the average value of the dispatch ratio of the corresponding ground agent in the historical data.

[0064] The flight delay penalty cost is the cost set when the flight is delayed. The product of the time length corresponding to the time adjustment range and the delay penalty coefficient is multiplied by the delay penalty coefficient, which can be set according to the priority of the corresponding flight. The calculation method of the flight cancellation penalty cost and the active delay penalty cost is the same as the flight delay penalty cost, except that the flight cancellation penalty cost is set according to the canceled flights, and the active delay penalty cost needs to check the difference between the delay time generated by the active delay and the upper limit of the allowed interactive delay.

[0065] When the flight is canceled, the corresponding cost value can be queried from the database, and the active delay penalty cost can be actively checked in the database to mark the upper limit of the allowed interactive delay and the value corresponding to the active delay.

[0066] The final departure balance difference cost is essentially a statistic of the dispatch ratio, which calculates the number of dispatched flights distributed in each time period. The sum of the squares of the difference between the number of dispatched flights in each time period and the average number of flights in all time periods is calculated. When the value is small, it means that the distribution of dispatched flights in each time period is relatively uniform, which meets the relevant constraint rules of the current flight dispatching.

[0067] When checking the penalty cost in each time period, the time adjustment penalty cost, flight delay penalty cost, ground handler adjustment difference penalty cost, flight cancellation penalty cost and active delay penalty cost are summed up according to the corresponding time period of the flight, and the departure balance difference cost is added to obtain the transfer cost of the current time period. Here, the summation will be standardized to remove the dimension, and the transferable cost in each time period is obtained. The larger the output transfer cost value, the longer the flight scheduling time in the corresponding time period, and the relative cost will be larger, which means that the flight scheduling in the corresponding time period will interfere with the global interference more seriously, and the flight combination list in each time period needs to be iteratively solved to make the relative cost lower, and a better adjustment time interval is selected.

[0068] The constraints mentioned in the current scheme can be explained by the parameters described in Tables 1-2.

[0069] Table 1. Standard parameter table

[0070]

[0071] In Table 1, a plurality of parameters are indicated, wherein the total capacity, the departure capacity, the arrival capacity and the adjustment ratio represent the maximum allowable values, these parameters are commonly used in flight scheduling scenarios, which will explain the specific situation of ground handling of flights; such as setting the minimum transit time according to the aircraft type level (A-E / F), explaining the minimum value of the runway interval of arrival and departure flights and departure flights, ensuring safe switching of runway occupancy, and explaining the runway mode corresponding to each runway. At this time, the scheduling ratio and the ground handler guarantee amount belong to the part of soft constraints, and the values will be marked according to the values of each time period. These values are only illustrative data of the current scheme, and in actual use, they will be adjusted according to different airports and the number of corresponding flights.

[0072] Table 2. Constraint rule table

[0073]

[0074] In Table 2, the normal rules that need to be met at present are explained, these constraint rules will be used for the setting and processing of the penalty factor, and the penalty factor is directed to the specific constraint rule. As for the X value explained in Table 2, it can be directly extracted from the corresponding item of Table 1, or directly extracted from the database; after completing the setting and calculation of the corresponding transfer cost, in step S5, these data will be rechecked, and the conditions described by the hard constraints will be rechecked to prevent the problem of combination failure after selecting the relevant flight data through the time gap.

[0075] In addition to determining the scheduled flights in each time period, the implementation of step S3 further includes: obtaining multiple batch flights of the same period, the same airport or the same runway respectively, determining whether there is range scheduling for the batch flights, and taking the on-time flight as the target element when there is range scheduling.

[0076] In the scenario of range scheduling, the same time can represent all departure flights within 15 minutes, and when 80% of the departure flights need to be scheduled, it means that there is range scheduling in the same period. Similarly, if at least 60% of the flights under the same airport code need to be scheduled, it means that the airport code personnel is short and needs to schedule their flights. Similarly, if at least 60% of the flights on the same runway need to be scheduled, it means that the weather conditions cause the runway to take off and land with reduced efficiency, and the schedule needs to be adjusted to ensure the safety of takeoff and landing. After meeting these proportions, it is considered that there is range scheduling for the corresponding flights.

[0077] In range scheduling, on-time flights need to be guaranteed. These flights may be delayed and adjusted due to batch scheduling of multiple flights, and the corresponding flights need to be recorded. These flights include transfer connection flights, international flights, large aircraft, etc., which are relatively large flights in flight reading and are easily affected by scheduling.

[0078] The pre- and post-sequences of the target element are obtained, and the corresponding flights are output to the scheduled flight list when the transfer cost of the corresponding pre- and post-sequences is minimized.

[0079] For range scheduling, the situation in step S3 for multiple flight scenarios will be further described. The on-time flights are recorded when scheduling, and multiple groups of flights are associated in the scenario of batch adjustment optimization, and these flights are synchronized to the list to determine the completeness of the data solution when the subsequent decomposition sub-constraints are solved, so that the overall cost of the calculation process is controllable.

[0080] In one embodiment of the present application, as shown in Figure 5 The implementation of step S4 includes: S41, based on the state definition and transfer cost corresponding to the scheduled flight list, the scheduled flight list is divided into multiple sub-constraints, and the sub-constraints are used for iterative solution.

[0081] S42, when the sub-constraints are iteratively solved, a Lagrange relaxation model is constructed, a sub-gradient method is used to update the Lagrange multiplier, and the optimal solution of each sub-constraint is obtained; and the intersection after iterative solution is regarded as the global iterative result.

[0082] S43, the global iterative result is mapped to the unit time to determine the adjustment time interval in each unit time, and the adjustment quantity is determined according to the adjustment time interval.

[0083] The overall goal of S4 is to decompose the global optimization problem of the timing flight list into solvable sub-problems, find the optimal solution through dual solution, and then convert it into specific scheduling instructions per unit time.

[0084] When splitting the constraints, the flights can be segmented according to the hard and soft constraints mentioned in the constraint rules, such as runway, ground station, unit time window, and transfer cost, according to the state definition of the timing flight list, and the global constraints are divided into multiple sub-constraint problems for solving.

[0085] When splitting by runway, such as runway 1 timing constraints, runway 2 timing constraints; when splitting by ground station, such as ground station 210 guarantee constraints, ground station 211 guarantee constraints, etc.; when splitting by unit time window, such as 7:00-7:15 timing constraints, 7:15-7:30 timing constraints; when splitting by flight type, such as transfer flight timing constraints, international flight timing constraints, etc. These sub-constraints represent the current need for iterative solution to obtain the adjustment time interval under the global optimal solution. The local optimal solution of each sub-constraint can be iteratively solved in sequence or in parallel using methods such as an upper limit of 100 iterations and a convergence threshold of 0.01.

[0086] When constructing the Lagrange relaxation model, the hard constraints are converted into cost items through the Lagrange multiplier, and its function form can be represented as ; wherein, represents the functional form of the Lagrange relaxation model, and represents the objective function that needs to be solved for the current sub-constraint, represents the value of the sub-constraint from one state to another state , which is specifically represented by the transfer cost of the timing amplitude and the delay cost; represents the hard constraint part existing in the constraint rule, which is set in a numerical form, with a value less than or equal to 0, and a Lagrange multiplier greater than or equal to 0 is introduced to obtain the Lagrange relaxation model under dual solution; the hard constraint represents the interval between consecutive departure flights, which is quantitatively solved according to the Lagrange multiplier , and the state described at this time explains the actual state of the corresponding flight in waiting, takeoff, etc.

[0087] After the next gradient method updates the Lagrange multiplier, the solution of the relaxation model gradually approaches the original problem solution that satisfies the hard constraint, and its formula is: ; wherein, represents the Lagrange multiplier of the kth iteration, represents the step size, which can be set as a constant, and the constant is divided by the number of iterations to obtain the corresponding step size; represents the Lagrange multiplier of the k+1th iteration, represents the value of the hard constraint part at state to state For each sub-constraint, the result of the calculation will traverse the states of all flights involved, and the path involved in the calculation will be taken as the local optimal solution of the corresponding sub-constraint when the cumulative value of the objective function of these flights at the current sub-constraint to be solved is the smallest.

[0088] For example: the optimal solution of runway 1 sub-constraint: flight A is adjusted by 5 minutes, flight B is adjusted by 3 minutes; the optimal solution of ground station 210 sub-constraint: flight A is adjusted by 5 minutes, flight C is adjusted by 2 minutes; then find the intersection that satisfies multiple sub-constraints, at this time flight A is adjusted by 5 minutes, which is equivalent to the intersection of sub-constraints, which is the global iteration result of the current demand.

[0089] Then map the global result to the minimum time granularity of flight scheduling, such as a 15-minute window, to get the result that flight A is adjusted by 5 minutes (originally 7:10→7:15), and then determine whether the flight after adjustment in the corresponding window meets the condition such as 1.5-minute departure runway interval time, at this time it can be considered as the relatively optimal adjustment time interval, finally count the total number of flights to be adjusted according to the unit time, so as to get the adjustment quantity of the current output.

[0090] Preferably, when step S42 is implemented, the implementation further includes: judging whether the sub-constraint solved by iteration has conflict, if there is conflict, fragmenting the previous flight corresponding to the current flight and determining the fragment identifier of each previous flight.

[0091] Obtaining the union set of the fragment identifiers of the previous flights to obtain a new list of flights to be adjusted.

[0092] Matching the fragment identifiers with the subsequent flights of the current flight respectively, and taking the matched data as the global iteration result of the solution.

[0093] And taking the adjustment time interval of the matched subsequent flight as the adjustment time interval of the unit time, and merging the adjustment time intervals to obtain the output adjustment quantity.

[0094] When the sub-constraint solved by iteration has conflict, it means that the intersection of the optimal solutions of the sub-constraints is empty, and it is necessary to avoid that the global result has no solution due to constraint conflict to ensure the feasibility of the scheduling scheme.

[0095] First, conflict detection and pre-sequencing flight slice identification, if the intersection of all sub-constraint optimal solution is empty, it is determined that there is a sub-constraint conflict; At this time, the pre-sequencing flight directly associated with the conflict flight is identified by sub-constraint solution, that is, different pre-sequencing flights are labeled with the label corresponding to a sub-constraint solution; For example, the pre-sequencing flights of conflict flight A have two solutions A1 and A2, A1 (support runway 1 solution: 7:15), A2 (support ground 210 solution: 7:20), and the slice identification is "slice 1: A1 (7:15)", "slice 2: A2 (7:20)".

[0096] Second, find the union set of all pre-sequencing flight slice identification, that is, include all pre-sequencing flights supporting different sub-constraint solutions; These pre-sequencing flights and their associated conflict flights are included in the new time adjustment flight list; For example, the union set is {A1, A2, A}, and the new list contains these three flights, which avoids excluding key flights due to conflict.

[0097] Then, slice identification and post-sequencing flight matching are performed to generate global results, each pre-sequencing flight slice identification is matched with the post-sequencing flight of the conflict flight to ensure the time connection between pre-sequencing and post-sequencing; For example, the post-sequencing flight B of conflict flight A, if matching slice 1 (A1→A is adjusted to 7:15), B needs to be adjusted to 7:18 (interval 3 minutes); If matching slice 2 (A2→A is adjusted to 7:20), B needs to be adjusted to 7:23 (interval 3 minutes); The part without constraint conflict after matching is output as the global iteration result.

[0098] Finally, the number of all time-adjusted flights in the time window is merged, such as merging A and B, to obtain the adjustment number of 2.

[0099] This process further avoids global no solution, guarantees the coherence of pre-sequencing and post-sequencing association, makes the adjustment flight schedule comply with normal constraint rules, and prevents chain delay caused by conflict processing.

[0100] As for the way of adjusting time interval combined with time slot allocation, the adjustment time interval is used to sort and arrange the flights according to the adjustment time interval, to determine the adjustment number in each unit time period.

[0101] The implementation of step S5 also includes: performing logical and or operation on the obtained adjustment number, synchronizing the adjustment number to the flight state diagram, finding the intersection of the current flight state diagram when there are the same index values in the flight state diagram, and regarding the combined data after intersection as the output adjustment scheme.

[0102] In step S5, the input of step S4 is continued, the corresponding adjustment quantity is received, and the data is synchronized to the flight status diagram, logical and or operation is performed using the preset constraint rules, if the preset constraint rules are met, the relevant data is updated, and after ensuring that there is no conflict with the preset constraint rules; the index value on each node is checked, that is, the corresponding parameters are checked, repeated versions are removed, and an adjustment scheme with repeated versions removed is obtained.

[0103] The logical and or operation is essentially to compare the calculated data with the conditions in the preset constraint rules, when the conditions are met, the value is 0, and the relevant data is updated, otherwise the value is 1, the number of constraints that do not meet the constraints is recorded, and this part of the data is re-input into the iterative solving process, to complete the global optimization of flight scheduling.

[0104] Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application, which are still covered by the protection scope of the present application.

Claims

1. A flight departure guarantee resource configuration scheduling optimization management method, characterized in that, The method comprises the following steps: S1, data extraction is performed according to current departure and arrival flights to obtain a flight schedule list; S2, a flight status graph is constructed in a pre-sequence and post-sequence manner according to the runway, runway sequence and time interval corresponding to the flight schedule list; S3, when the state of the flight status graph changes, the maximum capacity of each node in the flight status graph within a unit time period and the minimum adjustment time interval are taken as constraint conditions to determine the transition cost of each state in the flight status transition process, and each flight is sorted into a timing flight list according to the transition cost; S4, the timing flight list is used for iterative solving to determine the optimal adjustment time interval within a unit time, and the adjustment quantity within a unit time period is determined according to the adjustment time interval and time slot allocation; S5, when the adjustment quantity meets the preset constraint rule, the corresponding flights are combined, and the combined data is regarded as an output adjustment scheme; The implementation of step S1 comprises the following steps: S11, the number of flights to be dispatched at any time is obtained, and when the number of flights to be dispatched meets a dispatching-out ratio, the arrival and departure flight data are spliced, and the data start and end time interval at the time of splicing is recorded; the dispatching-out ratio refers to the ratio of the number of normally dispatched flights and the number of abnormally dispatched flights to the total number of flights, that is, the ratio of the number of normally dispatched flights and the number of abnormally canceled flights to the total number of flights; S12, the spliced flight data is removed according to the planned takeoff time and the actual takeoff time within the data start and end time interval, and is output as a flight schedule list; The implementation of step S3 comprises the following steps:

2. The flight departure guarantee resource configuration scheduling optimization management method according to claim 1, characterized in that, S31, a state change signal is generated in response to the change of the current flight status graph, and the state change signal is constructed according to the connection relationship of the current flight status graph; S32, each flight under state transition is read by logical operation to determine whether there is a conflict between the pre-sequence and post-sequence flights and the current flight, and a state transition matrix is constructed using the data after determination; S33, the timing amplitude, the dispatching-out ratio and the delay cost at each dispatching are taken as penalty factors to construct the transition cost of the state transition matrix in multiple time periods; S34, based on the transition cost of the state transition matrix, the transition probability meeting the constraint in each time period is calculated, and the dispatching-out ratio is output as a timing flight list in descending order according to the transition probability value; The implementation of step S4 comprises the following steps: S41, based on the state definition and transition cost corresponding to the timing flight list, the dispatching flight list is divided into multiple sub-constraints, and the sub-constraints are used for iterative solving; S42, when the sub-constraints are iteratively solved, a Lagrange relaxation model is constructed, a subgradient method is used to update the Lagrange multiplier, and the optimal solution of each sub-constraint is obtained; and the intersection after iterative solving is regarded as a global iterative result; S43, the global iterative result is mapped to a unit time to determine the adjustment time interval at each unit time, and the adjustment quantity is determined according to the adjustment time interval. The implementation of step S2 comprises the following steps: S21, judging whether there is a previous flight for the current flight, if there is, modeling the previous flight and the subsequent flight of the current flight, and determining the scheduling state of each flight according to the runway, runway sequence and time interval data of the previous flight and the subsequent flight; S22, if there is no previous flight, determining the scheduling state of each flight according to the runway, runway sequence and time interval corresponding to the subsequent flight of the current flight and other flights; S23, based on the scheduling state of each flight, taking the shortest time interval between the scheduling states as the target cycle time, and taking the scheduling mode of the target cycle time as the state change of the flight state diagram; The shortest time interval between the scheduling states refers to the time difference between the end state of the previous flight and the start state of the current flight, and the time difference between the end state of the current flight and the start state of the subsequent flight, which reflects the minimum interval of airport ground scheduling and resource recovery time. This value will focus on the process of resource reuse and release, and explain the time interval between the current flight and the next task after completing the task.

3. The flight departure guarantee resource configuration scheduling optimization management method according to claim 2, characterized in that, The implementation of step S23 further includes: According to the connection relationship between the previous flight and the current flight, a transfer time matrix composed of different previous flights and the current flight is generated; the minimum transfer time of different previous flights and the current flight is recorded in the transfer time matrix; The transfer time matrix is associated with the flight schedule list corresponding to the current flight, and the expected departure time of the outbound flight is updated.

4. The flight departure guarantee resource configuration scheduling optimization management method according to claim 3, characterized in that, The implementation of step S33 further includes: The time adjustment range is directed to the time adjustment penalty cost and the outbound balance difference cost, the time adjustment ratio is directed to the ground station time adjustment difference penalty cost, and the delay cost is directed to the flight delay penalty cost, the flight cancellation penalty cost and the active delay penalty cost. The sum of the penalty costs of each time period is taken as the transition cost of the state transition matrix.

5. The flight departure guarantee resource configuration scheduling optimization management method according to claim 1, characterized in that, The implementation of step S3 further includes: A plurality of batch flights of the same time period, the same ground or the same runway are obtained respectively, whether the batch flights exist range scheduling is determined, and the flight on time when scheduling is taken as the target element when the range scheduling exists; The previous flight and the subsequent flight of the target element are obtained, and the corresponding flights are output to the time adjustment flight list when the transition cost corresponding to the previous flight and the subsequent flight is the minimum.

6. The flight departure guarantee resource configuration scheduling optimization management method according to claim 5, characterized in that, When step S42 is implemented, the implementation thereof further includes: Judging whether the sub-constraint of the iterative solution exists conflict, if there is conflict, the previous flight corresponding to the current flight is identified by fragmentation, and the fragmentation identification of each previous flight is determined; The previous flights in the union set of the fragmentation identification are obtained, and a new time adjustment flight list is obtained; The fragmentation identification is matched with the subsequent flight of the current flight respectively, and the matched data is taken as the global iterative result of the solution; And the adjustment time interval corresponding to the matched subsequent flight is taken as the adjustment time interval of the unit time, and the adjustment quantity is obtained after merging the adjustment time interval.

7. The flight departure guarantee resource configuration scheduling optimization management method according to claim 1, characterized in that, The implementation of step S5 further includes: Logical and or operation is performed on the obtained adjustment quantity, and the adjustment quantity is synchronized to the flight state diagram. When there is the same index value in the flight state diagram, the intersection of the current flight state diagram is calculated, and the combined data after the intersection is taken as the output adjustment scheme.

Images