Terminal airspace flight scheduling decision-making method and system based on mixed linear integer programming

Abstract

The embodiment of the invention discloses a terminal airspace flight scheduling decision-making method and system based on mixed linear integer programming, and the method comprises the steps: obtaining the information of flights participating in scheduling in a time window from a flight information base, and obtaining the environment information from an operation environment information base; constructing a flight scheduling set; defining decision variables according to business requirements, and constructing a target function of a terminal airspace single-path flight ordering scheduling model; constructing a constraint condition library of the scheduling model; selecting a model according to business requirements; solving and calculating; and outputting decision information of the flight needing to be scheduled. According to the method, complex operation logic and control operation rules between flight pairs operated in a terminal airspace are accurately and abstractly modeled into mathematical linear constraint expression, and an objective function is constructed by minimizing the air waiting time of an incoming flight and minimizing the scheduling time deviation of the incoming and leaving flights; compared with a traditional first-arrival first-service algorithm, a flight sequence with higher efficiency can be obtained, and serious flight time slot waste cannot be generated.

Description

Technical Field

[0001] The present invention relates to the technical field of flight scheduling, and in particular to a terminal airspace flight scheduling decision method and system based on mixed linear integer programming. Background Art

[0002] With the continuous growth of air traffic demand, the contradiction between the current large-volume, strongly coupled air traffic flow and airspace resource capacity has become increasingly prominent. The traditional control and command method based on radar and procedural control and manual experience has led to many problems such as low operating efficiency and high workload of controllers. Terminal airspace flight scheduling aims to reasonably sort the terminal airspace aircraft queue in the time domain without violating the safe operating interval, and decide the optimal transit time, sequence, interval maintenance, speed control and other scheduling information for each flight, and assist controllers in formulating terminal airspace scheduling strategies. This technology can effectively increase the terminal airspace capacity under the existing operating resource conditions, alleviate airspace congestion, and reduce flights circling at high altitudes and waiting on the ground.

[0003] Prior art 1 is a multi-priority mode arrival sorting method and device thereof: This technology relates to a multi-priority mode arrival sorting method, firstly calculating the priority and sorting point of all flights to be sorted, the priority from high to low is manually frozen landing time, emergency secondary code, special plane VIP, priority route and ordinary priority; then based on the priority, sorting is combined with the interval between flights to determine the landing order of flights, calculate the arrival order of all flights, and establish an arrival flight sequence table. The disadvantage of this technology is that it is essentially based on the "first come, first served" scheduling strategy, that is, sorting in order of flight priority, and can only provide a set of feasible sorting strategies, which is very likely to deviate from the global optimal strategy.

[0004] Prior art 2 is a terminal area trailing interval approach management decision support system: this technology includes two parts: trailing interval management and time-based management strategies. Among them, the time-based management strategy allocates the time for each aircraft to arrive at the pre-specified waypoint; the trailing interval management strategy limits the required distance / time interval between consecutive aircraft. The disadvantages of this technology are: first, the constructed model is a nonlinear model, which increases the computational burden and is not conducive to the realization of hardware facilities; second, the patent ignores a variety of key decision factors in actual operation, such as flight operation speed and waiting mode decisions.

[0005] Prior art 3 is a method for fast sorting and optimizing the scheduling of incoming flights based on composite dispatch rules: this technology constructs a cost-based approach flight sorting model and designs composite dispatch rules to effectively obtain a good landing sequence; secondly, an optimization model is constructed based on the landing order, objective function, and constraints, and a solver is used to obtain the optimized landing time of the approaching flight. The disadvantage of this technology is that it only determines the time of the flight at the terminal corridor entrance point and the runway point, and has not yet made a more fine-grained decision on the time and order of the metering points passed by the terminal airspace flight.

[0006] Prior art 4 is a multi-performance optimization sorting method for approaching flights at a multi-runway airport: this technology allocates runways to flights and determines approach routes by defining a multi-performance runway allocation evaluation function, and optimizes the delay allocation process and determines the landing time of flights according to the comprehensive priority of the flights. The disadvantage of this technology is that the performance of the scheduling model is affected by the accuracy of the evaluation model, and when converting the correlation between the operating data into evaluation time, flow and punctuality factor, information loss may be encountered, which may lead to inaccurate or incomplete evaluation results, thereby affecting the overall performance. Summary of the invention

[0007] The present invention proposes a terminal airspace flight scheduling decision model and system based on hybrid linear integer programming, the purposes of which include:

[0008] Assist controllers to formulate flight scheduling strategies in terminal airspace to reduce control workload and achieve the transition from traditional manual experience-based decision-making to data-driven decision-making;

[0009] Reasonable planning and scheduling of incoming and outgoing flights can help improve terminal airspace capacity and airspace resource utilization under existing operating resource conditions, and effectively alleviate congestion in busy terminal areas;

[0010] By calculating the arrival order and assigned arrival time of flights at each metering point, and analyzing the minimum time domain safety interval that flights need to maintain when passing through each metering point, the ability to perceive future operational situations can be improved, and safety accidents or symptoms can be advanced from in-process control to pre-emptive prevention.

[0011] To achieve the above objectives, in a first aspect, an embodiment of the present invention provides a terminal airspace flight scheduling decision method based on hybrid linear integer programming, comprising:

[0012] Step 1: Obtain flight information participating in scheduling within the time window from the flight information database, and obtain environmental information from the operating environment information database;

[0013] Step 2: Determine whether rolling time domain control needs to be enabled. If not enabled, all flights participate in building a flight scheduling set; if enabled, flights within the rolling time domain step length participate in building a flight scheduling set;

[0014] Step 3: Define decision variables according to business requirements, and construct the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights;

[0015] Step 4: Construct a constraint library for the terminal airspace single-path flight scheduling model;

[0016] Step 5: Select Model 1, Model 2 or Model 3 based on business requirements; Model 1, Model 2 and Model 3 include different constraints in the constraint library;

[0017] Step 6: Solve and calculate the selected model according to the constraint condition library to obtain a calculation result;

[0018] Step 7: Outputting decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business requirements;

[0019] Step 8: Send the decision information to the route control seat, terminal control seat and tower control seat respectively.

[0020] Optionally, in step 3, when the rolling time domain control is enabled, additional rolling time domain constraints are added, including:

[0021] (1) Runway separation constraint of the previous frozen flight. If the previous frozen flight is on the same runway as the previous frozen flight, the minimum wake turbulence separation constraint is met;

[0022] (2) The minimum flight interval constraint of consecutive metering nodes of the previous frozen flight. If the previous frozen flight passes through the same metering point, the subsequent flight must ensure that the minimum flight time interval constraint is met;

[0023] (3) The consecutive metering nodes of the preceding frozen flights are constrained by uniqueness, ensuring that only one flight passes through the metering node first during the flight assignment process.

[0024] In the second aspect, the embodiment of the present application also provides a terminal airspace flight scheduling decision system based on hybrid linear integer programming, including a flight information database, an operating environment information database, a scheduling model construction module, a solution calculation module, a decision designation module and a situation information display terminal;

[0025] The scheduling model building module is used to:

[0026] Obtain the flight information participating in the scheduling within the time window from the flight information database, and obtain the environment information from the operation environment information database;

[0027] Based on the acquired flight information and environmental information, it is determined whether the rolling time domain control needs to be enabled. If not enabled, all flights participate in building the flight scheduling set; if enabled, flights within the rolling time domain step participate in building the flight scheduling set;

[0028] Define decision variables according to business needs, and build the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights;

[0029] Construct a constraint library for the terminal airspace single-path flight scheduling model;

[0030] Select Model 1, Model 2 or Model 3 based on business requirements; Model 1, Model 2 and Model 3 include different constraints in the constraint library;

[0031] The solution calculation module is used to solve the selected model according to the constraint condition library to obtain a calculation result;

[0032] The decision designation module is used to output the decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business requirements;

[0033] The situation information display terminal is used to display the decision information.

[0034] Furthermore, the system also includes an additional constraint building module, which is used to add rolling horizon additional constraint conditions when the rolling horizon control is enabled.

[0035] The technical solution implemented in the embodiment of the present invention is obviously different from the "first-come, first-served strategy". Specifically, this technology uses a hybrid linear integer programming model to accurately abstract and model the complex operating logic between flight operations and control operating rules into a series of mathematical linear constraint expressions, combined with the designed linear objective function, and through a global optimization algorithm or a mature industrial solver, the optimal scheduling strategy can be found in the entire domain.

[0036] In addition, compared with the prior art, the method proposed in the present invention has richer decision-making information. It not only includes the flight sequence and time in the traditional model, but also introduces decision contents such as flight operating speed, interval maintenance, and waiting mode. It can provide decision support services with higher reliability and flexibility in complex operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the drawings required for use in the specific implementation or the description of the prior art are briefly introduced below.

[0038] Figure 1It is a flow chart of a terminal airspace flight scheduling decision method based on hybrid linear integer programming provided by an embodiment of the present invention;

[0039] Figure 2 and Figure 3 yes Figure 1 A magnified view of the middle module;

[0040] Figure 4 yes Figure 1 Another flow chart of the method shown. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0043] Please refer to Figure 1 and Figure 4 , a terminal airspace flight scheduling decision method based on hybrid linear integer programming provided by an embodiment of the present invention includes:

[0044] Step 1: Obtain the flight information involved in the scheduling within the time window from the flight information database, and obtain the environment information from the operating environment information database.

[0045] Specifically, the flight information participating in the scheduling within the time window is obtained from the flight information database, including flight number, flight arrival and departure type, flight type (light, medium, heavy), flight runway allocation plan, arrival flight corridor entrance points, and arrival flight approach path information.

[0046] The following information is obtained from the operating environment information database: corridor entrance location information, approach route structure information, metering point location and type information, metering point waiting stack configuration, runway operation mode, and traffic flow management measures.

[0047] Manually output model parameter information, including but not limited to the maximum allowable delay for approaching and departing runways, the maximum allowable waiting time at the approach corridor entrance, the horizontal interval between consecutive flights in the terminal area (nautical miles), the allowable flight speed adjustment range, the wake turbulence interval of different types of flights, and the maximum capacity of the waiting stack.

[0048] Step 2: Determine whether rolling time domain control needs to be enabled. If not enabled, all flights participate in building a flight scheduling set; if enabled, flights within the rolling time domain step participate in building a flight scheduling set.

[0049] Specifically, the controller determines whether to enable rolling time domain control based on the number of flights to be scheduled.

[0050] Step 3: Define decision variables according to business requirements, and construct the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights.

[0051] Step 4: Construct a constraint library for the terminal airspace single-path flight sequencing and scheduling model.

[0052] The constraint library includes multiple constraints, including but not limited to:

[0053] (1) Time window constraints for incoming flights at the terminal boundary point, i.e. ensuring that the arrival time of incoming flights at the corridor entrance must be within the set time interval;

[0054] (2) The time window constraints for arrival and departure flights at the runway metering point, i.e. ensuring that the take-off or landing time of the arriving flight must be within the set time interval;

[0055] (3) Wake turbulence separation constraints for flights arriving and departing at the runway metering point, i.e., ensuring that the minimum wake turbulence separation interval must be met between consecutive flight pairs on the same runway;

[0056] (4) Terminal metering point flight speed range constraints, i.e., the maximum / minimum speed range of the flight at the terminal metering point;

[0057] (5) Approach metering point flight speed range constraints, i.e., the maximum / minimum speed range of the flight operating at the approach metering point;

[0058] (6) Terminal metering point approach flight transit time constraint, i.e., constraining the flight time of a flight passing through adjacent terminal metering nodes at the current speed;

[0059] (7) The time constraint of the approaching flight based on the waiting stack. If there is a waiting stack at the terminal metering point, the flight time of the flight passing through the adjacent terminal metering node at the current speed is constrained;

[0060] (8) Approach metering point approach flight transit time constraint, i.e., constraining the flight time of the flight passing through the adjacent approach metering node at the current speed;

[0061] (9) Intersection separation flight time interval expression, which quantifies the minimum separation time interval that any flight pair needs to meet based on the flight speed at the intersection measurement node ("T-shaped intersection" or "Y-shaped intersection");

[0062] (10) The expression of the flight time interval for continuous link separation is based on the flight speed of any flight pair at the same link measurement node to quantify the minimum separation time domain interval that needs to be met;

[0063] (11) The minimum separation interval constraint for consecutive flights at the common node ensures that any flight pair before and after must meet the minimum separation interval in the common metering node;

[0064] (12) The order of flights passing through each metering point is uniquely constrained. If there is a common metering node for a flight pair, ensure that only one flight passes through the metering node first during the flight assignment process;

[0065] (13) Each metering node prevents overtaking conflicts. Since the flight segment is unidirectional and unique, it ensures that there will be no overtaking between flight pairs during the flight.

[0066] (14) Overtaking constraints are allowed based on the metering point of the waiting stack. If the preceding flight has a waiting stack, the succeeding flight may overtake. Otherwise, overtaking will definitely not occur.

[0067] (15) Waiting stack waiting time range constraint, that is, ensuring that the waiting time of the flight at each terminal metering point is limited to a certain time range;

[0068] (16) The waiting stack capacity constraint ensures that the number of flights waiting simultaneously at each terminal metering point is less than or equal to the waiting stack capacity.

[0069] Step 5: If step 2 selects to enable rolling horizon control, add additional rolling horizon constraints.

[0070] Among them, additional constraints include:

[0071] (1) Runway separation constraint of the previous frozen flight. If the previous frozen flight is on the same runway as the previous frozen flight, the minimum wake turbulence separation constraint is met;

[0072] (2) The minimum flight interval constraint of consecutive metering nodes of the previous frozen flight. If the previous frozen flight passes through the same metering point, the subsequent flight must ensure that the minimum flight time interval constraint is met;

[0073] (3) The consecutive metering nodes of the preceding frozen flights are constrained by uniqueness, ensuring that only one flight passes through the metering node first during the flight assignment process.

[0074] Step 6: Select Model 1, Model 2 or Model 3 based on business requirements.

[0075] The scheduling models constructed above include model 1, model 2 or model 3. You can select the model according to your business needs:

[0076] Model 1 (constant speed and the metering point does not consider the waiting mode), the constraints are (1), (2), (3), (6), (8), (11), (12), (13) in the above step 4;

[0077] Model 2 (flight speed and interval are involved in decision making but the waiting mode is not considered), the constraints are (1), (2), (3), (4), (5), (6), (8), (9), (10), (11), (12), (13) in step 4 above;

[0078] Model 3 (flight speed and interval are involved in decision-making and the waiting mode is considered), the constraints are (1), (2), (3), (4), (5), (7), (8), (9), (10), (11), (12), (14), (15), and (16) in step 4 above.

[0079] Step 7: Solve and calculate the selected model according to the constraint library to obtain the calculation results.

[0080] In specific implementation, the solution calculation method can be selected from commercial solvers such as CPLEX / GURBOI, or group simulation algorithms such as genetic algorithms and bird flock algorithms, or traditional mathematical programming methods such as branch and bound and simplex method.

[0081] Step 8: Output the decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business needs.

[0082] Among them, the decision information output by model 1 includes the time when the flight passes each metering point and the order in which the flight passes the metering points;

[0083] The decision information output by model 2 includes the time when the flight passes each metering point, the order in which the flight passes the metering point, the speed value of the flight at each metering point, and the flight interval maintenance recommendation of the flight at each metering point in the time domain;

[0084] The decision information output by model 3 includes the time when the flight passes each metering point, the order in which the flight passes through the metering points, the speed value of the flight at each metering point, the flight time interval maintenance recommendation at each metering point, whether the flight executes the waiting procedure at each metering point, and the waiting time executed by the flight at each metering point.

[0085] Step 9: Send the decision information to the route control seat, terminal control seat and tower control seat respectively.

[0086] Furthermore, the mathematical expression of the above model is explained as follows:

[0087] According to the above steps, the mathematical expressions of the objective functions and constraints of each model are elaborated in detail.

[0088] According to step 1, the following data can be obtained, assuming:

[0089] Fa is the set of arriving flights;

[0090] Fd is the departure flight set;

[0091] CTA f Plan arrival times for flights;

[0092] CTD f Plan departure times for flights;

[0093] TMAtime f It is the time when the incoming flight arrives at the terminal corridor entrance;

[0094] u is the edge set of the terminal airspace topology structure; u f is the edge set of flight f following the standard approach procedure; is the terminal edge set of flight f according to the standard approach procedure; is the set of approach edges for flight f according to the standard approach procedure;

[0095] v is the set of metering nodes in the terminal airspace topology structure, is the set of measurement points for runway f of flight, is the terminal metering point set of flight f, is the set of final approach metering points for flight f, v f is the set of all measurement points of the standard approach procedure for flight f, v runway is the set of all runway points, v enter Gather at all corridor entrances.

[0096] d g,v is the Euclidean distance length of edge (g, v);

[0097] H max / H min The maximum / minimum waiting time of a flight in the waiting stack;

[0098] Cap Hold The maximum waiting capacity of the waiting stack;

[0099] Manual input of model parameters:

[0100] The default horizontal separation requirement for flights, usually set to 3NM;

[0101] E max The maximum allowed waiting time at the entrance of the corridor;

[0102] D max The maximum permissible delay time for approaching and departing runways;

[0103] d f,f' is the wake interval between the preceding flight f and the succeeding flight f';

[0104] Vec avg is the historical nominal speed value of the flight;

[0105] is the maximum speed of flight f passing through the measuring point v;

[0106] is the minimum speed of flight f passing through the measuring point v;

[0107] M is a very large constant, usually 10 5 ;

[0108] (I) Model-Mathematical Expression

[0109] For the mathematical expression of model 1 in step 6 above, define the following decision variables:

[0110] 1) Decision variables for arrival flight transit time: is the arrival time of the incoming flight f at node v, is the arrival time of the incoming flight f at the runway node;

[0111] 2) Departure time decision variables of departing flights: is the departure time of departing flight f at the runway point;

[0112] 3) Public node flight order decision variable: Λ f,f',v is a binary 0-1 decision variable, which means that if flight f arrives at the common node v before flight f', the value is 1, otherwise it is 0;

[0113] According to step 3, the objective function (1) of model 1 is set to minimize the deviation of the flight schedule time of arrival and departure, which can be expressed as follows:

[0114]

[0115] The objective function (2) is set to minimize the waiting time of the incoming flight, which can be expressed as follows:

[0116]

[0117] Therefore, the overall objective function of model 1 is expressed as:

[0118] OBJ=min(α·OBJ1+β·OBJ2)

[0119] Wherein, α and β are weight factors, satisfying α+β=1, and α, β>1.

[0120] The constraints of model 1 are the constraints (1), (2), (3), (6), (8), (11), (12), and (13) in step 4 above. For model 1, constraint (1) is the time window constraint for the arrival flight at the terminal boundary point, which can be expressed as follows:

[0121]

[0122]

[0123] For the first constraint condition (2) of the model, the time window constraint of the flight arrival and departure at the runway metering point can be expressed as follows:

[0124]

[0125] The constraint condition (3) of model 1 is the wake turbulence interval constraint of the flight arriving and departing at the runway metering point, which can be expressed as follows:

[0126]

[0127] The constraint condition (6) of model 1 is the time constraint of the arrival flight at the terminal metering point, which can be expressed as follows:

[0128]

[0129] The constraint condition (8) of model 1 is the approach metering point arrival flight transit time constraint, which can be expressed as follows:

[0130]

[0131] For model 1, constraint condition (11) is the minimum separation flight interval constraint of common nodes of consecutive flights (including terminal nodes and approach nodes), which can be expressed as follows:

[0132]

[0133] For the first constraint (12) of the model, the flight sequence of each metering point passes the unique constraint, which can be expressed as follows:

[0134]

[0135] The constraint condition (13) of model 1 is the constraint to prevent overtaking conflicts at each metering node, which can be expressed as follows:

[0136]

[0137] After the model is solved in step 7 above, step 8 outputs the decision information including the time when the flight passes each metering point and the order in which flights pass through the metering pointsf,f',v .

[0138] (II) Mathematical expression of model 2

[0139] For the mathematical expression of Model 2 in step 6 above, define the following decision variables:

[0140] 1) Decision variables for arrival flight transit time: is the arrival time of the incoming flight f at node v, is the arrival time of the incoming flight f at the runway node;

[0141] 2) Departure time decision variables of departing flights: is the departure time of departing flight f at the runway point;

[0142] 3) Public node flight order decision variable: Λ f,f',v is a binary 0-1 decision variable, which means that if flight f arrives at the common node v before flight f', the value is 1, otherwise it is 0;

[0143] 4) Decision variables for the speed of approaching flights passing through each metering point: Dec f,v It is the inverse value of the speed of flight f passing through the measuring point v.

[0144] 5) Minimum flight time interval constraints for flights passing through a common metering point: SepMeter f,f',v It is the minimum flight time interval between the preceding flight f and the succeeding flight f' passing through the common metering point v.

[0145] 6) Linearization auxiliary decision variables: w0, w1, w2 are binary 0-1 auxiliary decision variables, which assist in converting nonlinear equations into linear equations and have no special physical meaning;

[0146] The objective function of model 2 has the same mathematical expression as that of model 1. Please refer to the above description.

[0147] The constraints of model 2 are the constraints (1), (2), (3), (4), (5), (6), (8), (9), (10), (11), (12), (13) in step 4 above; among them, the expressions of constraints (1), (2), (3), (12), (13) are the same as those of model 1 and will not be repeated here. The following will explain the constraint expressions that are different from model 1:

[0148] For model 2, constraint condition (4) is the flight speed range constraint of the terminal metering point. At the same time, the flight maintains a constant speed or deceleration during the approach process in the terminal airspace, which can be expressed as follows:

[0149]

[0150] The constraint condition (5) of model 2 is the flight speed range constraint of the approach metering point, which can be expressed as follows:

[0151]

[0152] For model 2, constraint condition (6) is the transit time constraint of the terminal metering point arrival flight, which can be expressed as follows:

[0153]

[0154] The constraint condition (8) of model 2 is the approach metering point arrival flight transit time constraint, which can be expressed as follows:

[0155]

[0156] The constraint condition (9) of model 2 is expressed as the time interval of the intersection separation flight, which can be expressed as follows:

[0157] When Dec f,v ≤Dec f',v Then the interval satisfies

[0158] When Dec f,v >Dec f',v Then the interval satisfies

[0159]

[0160] The above constraint condition (9) can be expressed linearly as:

[0161] Dec f,v ≥Dec f',g' -M(1-w0)

[0162] Dec f,v ≤Dec f',g' +w0M

[0163]

[0164] The constraint condition (10) of model 2 is expressed as the flight time interval of continuous link separation, which can be expressed as follows:

[0165]

[0166] The above constraint (10) can be expressed in linear form:

[0167]

[0168] For the second model constraint (11), which is the minimum separation interval constraint for common nodes of consecutive flights, the SepMeter calculated based on the above constraints (9)-(10) is f,f',v , the minimum separation flight interval of common nodes of consecutive flights can be expressed as follows:

[0169]

[0170] After the model 2 is solved in step 7 above, step 8 outputs the decision information including the time when the flight passes each metering point Flights pass through metering points in order Λ f,f',v 、Flight speed value at each measurement point 1 / Dec f,v , flights at each metering point flights to maintain time domain flight interval recommendations SepMeter f,f',v .

[0171] (III) Mathematical expression of model 3

[0172] For the mathematical expression of model 3 in step 6 above, define the following decision variables:

[0173] 1) Decision variables for arrival flight transit time: is the arrival time of the incoming flight f at node v, is the arrival time of the incoming flight f at the runway node;

[0174] 2) Departure time decision variables of departing flights: is the departure time of departing flight f at the runway point;

[0175] 3) Public node flight order decision variable: Λ f,f',v is a binary 0-1 decision variable, which means that if flight f arrives at the common node v before flight f', the value is 1, otherwise it is 0;

[0176] 4) Decision variables for the speed of approaching flights passing through each metering point: Dec f,v It is the inverse value of the speed of flight f passing through the measuring point v.

[0177] 5) Minimum flight time interval variable for flights passing through a common metering point: SepMeter f,f',v It is the minimum flight time interval between the preceding flight f and the succeeding flight f' passing through the common metering point v.

[0178] 6) Flight waiting time variable at the terminal metering point: H f,g,v When selecting flight f from the previous node g to the node v, the waiting time required for the waiting stack at the node v;

[0179] 7) Flight waiting decision variables at the terminal metering point: ξ f,vIt is a binary 0-1 decision variable. When flight f is selected from the previous node to the current node v, it is 1 if waiting occurs in the waiting stack of node v, otherwise it is 0;

[0180] 8) Flight interval linearization auxiliary decision variables: w0, w1, w2 are binary 0-1 auxiliary decision variables, which assist in converting nonlinear equations into linear equations and have no special physical meaning;

[0181] 9) Waiting stack capacity linearization auxiliary decision variable: B f,f',v is a binary 0-1 auxiliary decision variable. , the value is 1, otherwise it is 0.

[0182] The objective function of model 3 is different from that of models 1 and 2. The objective function (1) of model 1 is set to minimize the deviation of the flight schedule time of arrival and departure, which can be expressed as follows:

[0183]

[0184] The objective function (2) is set to minimize the waiting time of the incoming flight, which can be expressed as follows:

[0185]

[0186] in, The following constraints are met:

[0187]

[0188] Therefore, the overall objective function of model three is expressed as:

[0189] OBJ=min(α·OBJ1+β·OBJ2)

[0190] Wherein, α and β are weight factors, satisfying α+β=1, and α, β>1.

[0191] The constraints of model three are the constraints (1), (2), (3), (4), (5), (7), (8), (9), (10), (11), (12), (14), (15), and (16) in step 4 above. Among them, constraints (1), (2), (3), (4), (5), (8), (9), (10), (11), and (12) are the same as those of model two and will not be repeated here.

[0192] The following is an explanation of the constraint expression that is different from model 2:

[0193] The third constraint (7) of the model is the arrival flight transit time constraint based on the waiting stack, which can be expressed as follows:

[0194]

[0195] The model three constraint condition (14) is the overtaking constraint based on the metering point of the waiting stack, which can be expressed as follows:

[0196]

[0197] The third constraint (15) of the model is the waiting time range constraint of the waiting stack, which can be expressed as follows:

[0198]

[0199] The third constraint (16) of the model is the waiting capacity constraint of the waiting stack, which can be expressed as follows:

[0200]

[0201] in, is the indicator function, considering the maximum capacity of the waiting stack of the control load as Cap Hold , when the subsequent flight's transit time meets this interval, it is 1, otherwise it is 0. The constraint condition (16) is linearized as follows:

[0202]

[0203] After the model 3 is solved in step 7 above, step 8 outputs the decision information including the time when the flight passes each metering point Flights pass through metering points in order Λ f,f',v 、Flight speed value at each measurement point 1 / Dec f,v , flights at each metering point flights to maintain time domain flight interval recommendations SepMeter f,f',v , the waiting time of the flight at each measuring point

[0204] Based on the mathematical expression of the above models, the mixed linear integer programming form of each model is described below:

[0205] (1) Model 1 mixed linear integer programming form:

[0206] Objective Function

[0207]

[0208] Constraints:

[0209]

[0210] (2) Model 2 mixed linear integer programming form: Objective function:

[0211]

[0212] Constraints:

[0213]

[0214] ω1+ω2≥1

[0215] (3) Model 3 mixed linear integer programming form:

[0216] Objective Function

[0217]

[0218] Constraints:

[0219]

[0220] ω1+ω2≥1

[0221]

[0222] From the above description, it can be seen that the terminal airspace flight scheduling decision method based on hybrid linear integer programming provided by the embodiment of the present invention has the following advantages:

[0223] (1) The present invention accurately abstracts and models the complex operation logic and control operation rules between flight pairs operating in the terminal airspace as mathematical linear constraint expressions, and constructs an objective function by minimizing the air waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights. Compared with the traditional "first-come, first-served algorithm", a more efficient flight sequence can be obtained without causing serious waste of flight time slots.

[0224] (2) The objective function and constraints constructed by the present invention contain both integer variables and continuous variables, and are both linearized objectives and constraints. Therefore, the terminal airspace flight scheduling decision model can be converted into a mixed integer linear programming model. Compared with other patented nonlinear models, the biggest advantage is that the optimal solution can be searched in the entire domain, and mature industrial solvers on the market can be used to improve the solution quality.

[0225] (3) The present invention combines the complexity of actual operation scenarios and fully considers the key decision factors in actual operation that are ignored by traditional scheduling schemes, making scheduling decisions closer to the actual operation scheduling scenario requirements, thereby optimizing time slot resources more accurately and avoiding potential conflicts, time slot waste or scheduling bottlenecks in traditional scheduling methods. At the same time, multiple decision suggestions enable controllers to make scheduling decisions quickly and accurately, improving the scientificity and practicality of the scheduling scheme.

[0226] (4) The present invention relates to a mathematical construction method for multiple scheduling models, which can extract objective functions, constraints and solution algorithms corresponding to business scenarios, achieve rapid response and support for business scenario scheduling needs, improve model reusability, and improve the integration and selection convenience of scheduling solutions.

[0227] (5) The present invention also introduces rolling time domain control to establish a planning model with a forward-looking time window. By dividing a large flight data set into several scheduled flight subsets, rolling local optimization is performed repeatedly to decompose the global optimization problem into multiple local optimization problems for processing. This step not only simplifies the complexity of the problem, but also can approach the global optimal solution to a certain extent, thereby ensuring the calculation efficiency while also ensuring the optimization effect. At the same time, the use of rolling time domain control can be dynamically adjusted according to real-time status and demand changes, and the plan can be adjusted in time to deal with unexpected events and changes. This dynamic adjustment capability enables the model to better cope with uncertainties and changes in the system and improve the flexibility and adaptability of the scheduling plan.

[0228] Based on the same inventive concept, an embodiment of the present invention also provides a terminal airspace flight scheduling decision system based on mixed linear integer programming, including a flight information database, an operating environment information database, a scheduling model construction module, a solution calculation module, a decision designation module and a situation information display terminal;

[0229] The scheduling model building module is used to:

[0230] Obtain the flight information participating in the scheduling within the time window from the flight information database, and obtain the environment information from the operation environment information database;

[0231] Based on the acquired flight information and environmental information, it is determined whether the rolling time domain control needs to be enabled. If not enabled, all flights participate in building the flight scheduling set; if enabled, flights within the rolling time domain step participate in building the flight scheduling set;

[0232] Define decision variables according to business needs, and build the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights;

[0233] Construct a constraint library for the terminal airspace single-path flight scheduling model;

[0234] Select Model 1, Model 2 or Model 3 based on business requirements; Model 1, Model 2 and Model 3 include different constraints in the constraint library;

[0235] The solution calculation module is used to solve the selected model according to the constraint condition library to obtain a calculation result;

[0236] The decision designation module is used to output the decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business requirements;

[0237] The situation information display terminal is used to display the decision information.

[0238] Furthermore, the system also includes an additional constraint building module, which is used to add rolling horizon additional constraint conditions when the rolling horizon control is enabled.

[0239] It should be noted that for more specific working procedures of the system, please refer to the aforementioned method embodiment section, which will not be repeated here.

[0240] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A terminal airspace flight scheduling decision method based on hybrid linear integer programming, characterized in that: include: Step 1: Obtain flight information participating in scheduling within the time window from the flight information database, and obtain environmental information from the operating environment information database; Step 2: Determine whether rolling time domain control needs to be enabled. If not, all flights participate in building a flight scheduling set. If enabled, flights within the rolling time domain step will participate in building the flight scheduling set; Step 3: Define decision variables according to business requirements, and build the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights; Step 4: Construct a constraint library for the terminal airspace single-path flight scheduling model; Step 5: Select Model 1, Model 2 or Model 3 based on business requirements; Model 1, Model 2 and Model 3 include different constraints in the constraint library; Step 6: Solve and calculate the selected model according to the constraint condition library to obtain a calculation result; Step 7: Outputting decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business requirements; Step 8: Send the decision information to the route control seat, terminal control seat and tower control seat respectively.

2. The method according to claim 1, characterized in that In step 3, when rolling horizon control is enabled, additional rolling horizon constraints are added, including: (1) Runway separation constraint of the previous frozen flight. If the previous frozen flight is on the same runway as the previous frozen flight, the minimum wake turbulence separation constraint is met; (2) The minimum flight interval constraint of consecutive metering nodes of the previous frozen flight. If the previous frozen flight passes through the same metering point, the subsequent flight must ensure that the minimum flight time interval constraint is met; (3) The consecutive metering nodes of the preceding frozen flights are constrained by uniqueness, ensuring that only one flight passes through the metering node first during the flight assignment process.

3. The method according to claim 1, characterized in that The constraints in the constraint library include: (1) Time window constraints for incoming flights at terminal boundary points; (2) Time window constraints for arrival and departure flights at runway metering points; (3) Wake turbulence separation constraints for flights arriving and departing at the runway metering point; (4) Flight speed range constraints at terminal metering points; (5) Flight speed range constraints at approach metering points; (6) Time constraints for incoming flights at the terminal metering point; (7) The time constraint of the approaching flight based on the waiting stack. If there is a waiting stack at the terminal metering point, the flight time of the flight passing through the adjacent terminal metering node at the current speed is constrained; (8) Time constraints for approaching flight metering points; (9) The expression of the separation time interval at the intersection is to quantify the minimum separation time interval that any flight pair needs to meet based on the flight speed at the intersection measurement node; (10) The expression of the flight time interval for continuous link separation is based on the flight speed of any flight pair at the same link measurement node to quantify the minimum separation time domain interval that needs to be met; (11) Minimum separation interval constraints for common nodes of consecutive flights; (12) The flight sequence of each metering point passes the uniqueness constraint; (13) Constraints at each metering node to prevent overtaking conflicts; (14) Overtaking constraints are allowed based on the metering points of the waiting stack; (15) Waiting stack waiting time range constraint; (16) The waiting stack waits for the capacity-limited constraint.

4. The method according to claim 1, characterized in that Model 1 includes constraints (1), (2), (3), (6), (8), (11), (12), and (13); Model 2 includes constraints (1), (2), (3), (4), (5), (6), (8), (9), (10), (11), (12), and (13); Model 3 includes constraints (1), (2), (3), (4), (5), (7), (8), (9), (10), (11), (12), (14), (15), and (16).

5. The method according to claim 1, characterized in that For the mathematical expression of model 1 in step 5, define the following decision variables: 1) Decision variables for arrival flight transit time: is the arrival time of the incoming flight f at node v, is the arrival time of the incoming flight f at the runway node; 2) Departure time decision variables of departing flights: is the departure time of departing flight f at the runway point; 3) Public node flight order decision variable: Λ f,f',v is a binary 0-1 decision variable, which means that if flight f arrives at the common node v before flight f', the value is 1, otherwise it is 0; According to step 3, the objective function (1) of model 1 is set to minimize the deviation of the flight schedule time of arrival and departure, which can be expressed as follows: The objective function (2) is set to minimize the waiting time of the incoming flight, which can be expressed as follows: Therefore, the overall objective function of model 1 is expressed as: OBJ=min(α·OBJ1+β·OBJ2) Wherein, α and β are weight factors, satisfying α+β=1, and α, β>1.

6. The method according to claim 1, characterized in that For the mathematical expression of model 3 in step 5, define the following decision variables: 1) Decision variables for arrival flight transit time: is the arrival time of the incoming flight f at node v, is the arrival time of the incoming flight f at the runway node; 2) Departure time decision variables of departing flights: is the departure time of departing flight f at the runway point; 3) Public node flight order decision variable: Λ f,f',v is a binary 0-1 decision variable, which means that if flight f arrives at the common node v before flight f', the value is 1, otherwise it is 0; 4) Decision variables for the speed of approaching flights passing through each metering point: Dec f,v is the inverse value of the speed of flight f passing through the measuring point v; 5) Minimum flight time interval variable for flights passing through a common metering point: SepMeter f,f',v is the minimum flight time interval between the preceding flight f and the succeeding flight f' passing through the common metering point v; 6) Flight waiting time variable at the terminal metering point: H f,g,v When selecting flight f from the previous node g to the node v, the waiting time required for the waiting stack at the node v; 7) Flight waiting decision variables at the terminal metering point: ξ f,v It is a binary 0-1 decision variable. When flight f is selected from the previous node to the current node v, it is 1 if waiting occurs in the waiting stack of node v, otherwise it is 0; 8) Flight interval linearization auxiliary decision variables: w0, w1, w2 are binary 0-1 auxiliary decision variables, which assist in converting nonlinear equations into linear equations and have no special physical meaning; 9) Waiting stack capacity linearization auxiliary decision variable: B f,f',v is a binary 0-1 auxiliary decision variable. When , the value is 1, otherwise it is 0; The objective function (1) of model 1 is set to minimize the deviation of the flight schedule time of arrival and departure, which can be expressed as follows: The objective function (2) is set to minimize the waiting time of the incoming flight, which can be expressed as follows: in, The following constraints are met: Therefore, the overall objective function of model three is expressed as: OBJ=min(α·OBJ1+β·OBJ2) Wherein, α and β are weight factors, satisfying α+β=1, and α, β>1.

7. A terminal airspace flight scheduling decision system based on hybrid linear integer programming, characterized in that: It includes flight information database, operating environment information database, scheduling model building module, solution calculation module, decision-making designation module and situation information display terminal; The scheduling model building module is used to: Obtain the flight information participating in the scheduling within the time window from the flight information database, and obtain the environment information from the operation environment information database; Determine whether rolling time domain control needs to be enabled. If not, all flights participate in building the flight scheduling set. If enabled, flights within the rolling time domain step will participate in building the flight scheduling set; Define decision variables according to business needs, and build the objective function of the terminal airspace single-path flight scheduling model based on minimizing the waiting time of incoming flights and minimizing the scheduling time deviation of incoming and outgoing flights; Construct a constraint library for the terminal airspace single-path flight scheduling model; Select Model 1, Model 2 or Model 3 based on business requirements; Model 1, Model 2 and Model 3 include different constraints in the constraint library; The solution calculation module is used to solve the selected model according to the constraint condition library to obtain a calculation result; The decision designation module is used to output the decision information of the flights to be scheduled in the flight scheduling set according to the model and calculation results corresponding to the business requirements; The situation information display terminal is used to display the decision information.

8. The system according to claim 7, characterized in that The system also includes an additional constraint building module for adding additional constraints of the rolling horizon when the rolling horizon control is enabled, including: (1) Runway separation constraint of the previous frozen flight. If the previous frozen flight is on the same runway as the previous frozen flight, the minimum wake turbulence separation constraint is met; (2) The minimum flight interval constraint of consecutive metering nodes of the previous frozen flight. If the previous frozen flight passes through the same metering point, the subsequent flight must ensure that the minimum flight time interval constraint is met; (3) The consecutive metering nodes of the preceding frozen flights are constrained by uniqueness, ensuring that only one flight passes through the metering node first during the flight assignment process.

Images