A large airport runway scheduling method fusing rule triggering and genetic algorithm

Abstract

This invention discloses a large-scale airport runway scheduling method integrating rule-based triggering and genetic algorithms. The method includes: acquiring the airport's standard taxiway path map and all data of arriving and departing aircraft within the same time period, and preprocessing them; using a heuristic rule algorithm to sort the priority sequence of departing aircraft in each data set, obtaining a priority ranking for each departing aircraft; assigning runways to each departing aircraft according to their priority ranking, and setting corresponding start taxiing times; the resulting allocation is the scheduling scheme; using the generated scheduling scheme as the initial solution, generating a population based on the initial solution, and iteratively calculating using a genetic algorithm to finally obtain the optimal individual, which is the final scheduling scheme. This invention can further reduce the number of conflict points and flight delays during aircraft taxiing, improving airport safety and flight punctuality.

Description

Technical Field

[0001] This invention relates to the field of runway scheduling technology, and in particular to a large-scale airport runway scheduling method that integrates rule triggering and genetic algorithms. Background Technology

[0002] Currently, my country's economy is booming, leading to a surge in demand for air transport and a continuous increase in the number of aircraft. Aircraft activities primarily focus on two stages: flight and taxiing. It's worth noting that runways are exclusive, allowing only one aircraft to use them at a time; taxiway systems, however, can accommodate multiple aircraft simultaneously. However, with the increasing number of taxiways, airport ground layouts are becoming more complex, offering numerous taxiway route options for aircraft. However, airport ground aircraft taxiways are generally selected by dispatchers and pilots based on historical experience. This means that the existing airport resource allocation cannot be fully utilized, increasing both operational costs and the risk of aircraft conflicts during taxiing. Compared to economic factors such as taxiing distance, airports prioritize the number of conflict points and flight delays. They are willing to incur slightly higher costs during taxiing to minimize conflicts and delays. Therefore, developing appropriate taxiing plans to ensure the safe and efficient taxiing of aircraft has become a crucial task in airport management.

[0003] 1. Existing research primarily focuses on addressing conflict avoidance problems by altering factors influencing aircraft taxiing costs, such as path length and taxiing speed, and then finding the optimal path based on these changes. However, it rarely considers the impact of the initial taxiing time on the taxiing process. In other words, different initial taxiing times lead to different potential conflict scenarios, thus affecting the final number of aircraft conflicts, the number of delayed flights, and the total taxiing time. The waiting time for aircraft to avoid conflict is also a component of taxiing costs. Separating the initial taxiing time from runway scheduling settings does not effectively reduce aircraft taxiing costs or the risk of conflict during taxiing.

[0004] Therefore, when analyzing aircraft path conflict avoidance, the impact of changes in the initial taxiing time delay on potential conflicts in the future path should be considered. However, current modeling research that comprehensively considers the trade-off between taxiing costs and delay costs is still insufficient. Summary of the Invention

[0005] The purpose of this invention is to provide a large-scale airport runway scheduling method that integrates rule-based triggering and genetic algorithms, with the goal of reducing the number of conflict points and delayed aircraft during aircraft taxiing. By rationally arranging the initial taxiing time of aircraft and the corresponding runway selected, the solution efficiency of the algorithm is improved, thereby reducing the number of conflicts and delays during aircraft operation.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] A method for large-scale airport runway scheduling that integrates rule-based triggering and genetic algorithms includes:

[0008] Obtain the airport's standard taxiway map and all data of airport arrivals and departures within the same time period, and preprocess them.

[0009] The priority sequence of departing aircraft is sorted for each set of data using a heuristic rule algorithm to obtain the priority ranking of each departing aircraft.

[0010] Prioritize each departing aircraft, allocate runways to departing aircraft according to the priority, and set the corresponding start taxiing time. The situation after allocation is used as the scheduling plan.

[0011] The generated scheduling scheme is used as the initial solution. A population is generated based on the initial solution, and iterative calculations are performed through a genetic algorithm. Finally, the optimal individual is obtained through the genetic algorithm, which is the final scheduling scheme.

[0012] Furthermore, the preprocessing involves removing irrelevant data; the preprocessed data includes:

[0013] Departing aircraft data: parking position, assigned runway, aircraft type, aircraft departure time, and ground taxiing time; Arriving aircraft data: parking position, landing runway, and landing time.

[0014] Furthermore, the aircraft type, takeoff time, and ground taxiing time of each aircraft in the departing aircraft data are used as a set of data as input to the heuristic rule algorithm; the parking position, runway gate, and landing time of each aircraft in the arriving aircraft data are used as a set of data as input to conflict judgment.

[0015] Furthermore, the heuristic rule algorithm obtains the priority ranking of departing aircraft using the following formula:

[0016]

[0017]

[0018] Among them, I j(t,l) represents the priority index of the next departing aircraft operation j after operation l is completed; j is the operation sequence number, indicating the process of the departing aircraft with the corresponding sequence number proceeding to the corresponding runway gate according to the standard taxiing path from the start taxiing time and then taking off for departure; l represents the index of the completed operation; w j p represents the weight of task j. j This represents the ground taxiing time of the departing aircraft corresponding to task j;

[0019] d j Let t represent the departure time of the departing aircraft corresponding to task j, and t represent the current time. This indicates the average ground taxiing time for departing aircraft. s represents the average safe time interval between departing aircraft. lj This represents the safe time interval between job j and job l, where k1 and k2 are the look-ahead and scaling parameters, respectively; τ is the compactness factor, R is the range factor, η is the set-time severity factor, and n is the number of jobs. d represents the average takeoff time of departing aircraft, β represents the time correlation coefficient of variation, and d represents the average takeoff time of departing aircraft. max d min These represent the departure times of the latest and earliest departing aircraft, respectively.

[0020] Priority index I j Sort (t,l) from largest to smallest, which is the priority sorting of the corresponding departing aircraft.

[0021] Furthermore, the allocation of runways to departing aircraft according to their order includes:

[0022] First, a set of departing aircraft is created, containing relevant information for all departing aircraft to be sorted. Priority indices for each departing aircraft are obtained using a heuristic algorithm, and then sorted. The departing aircraft with the highest priority index is added to the planned aircraft set, then removed from the original departing aircraft set, and its current time is changed to the current time plus the aircraft's ground taxiing time. The priority indices of the remaining departing aircraft in the departing aircraft set are recalculated until all departing aircraft are placed into the planned aircraft set in priority order. Then, runway allocation begins: first, a runway is allocated to the departing aircraft with the highest priority index. If a runway is available (i.e., no aircraft is ready to take off), it is allocated to the nearest available runway according to the standard taxiway path diagram. Then, runway allocation is performed for the next highest priority departing aircraft. If no runway is available, allocation is performed only after the time interval between the aircraft and the aircraft on a non-available runway exceeds the safe time interval. This process is repeated until all departing aircraft are allocated their corresponding runways, completing the generation of the scheduling plan.

[0023] Furthermore, the initialization process of the genetic algorithm includes:

[0024] A set of aircraft taxiing schemes is initialized based on the initial solution, which serves as the population for the genetic algorithm. The initial solution takes the form of {starting taxiing time 1, runway assigned to departing aircraft 1; starting taxiing time 2, runway assigned to departing aircraft 2; ... starting taxiing time n, runway assigned to departing aircraft n}.

[0025] Based on the initial solution, the starting taxiing time and assigned runway of any number of departing aircraft are randomly changed. A new scheduling scheme is generated after each change. After a certain number of changes, all the new scheduling schemes and the initial solution are used as individuals in the initial population.

[0026] The method for each change is as follows:

[0027] The departure aircraft will be randomly selected at any time within 15 minutes before its departure time as its new starting taxi time; and one of the remaining runways in the same direction as the departure aircraft's assigned runway will be randomly selected as its new runway.

[0028] Furthermore, the selection operation of the genetic algorithm includes:

[0029] The selection operation uses the roulette wheel selection method. The roulette wheel is divided into population sizes based on the proportion of individual fitness values, and the size of the roulette wheel area is directly proportional to the fitness of an individual. The fitness is calculated as follows:

[0030] F=ω1g1(c1,c2,…,c 2i )+ω2g2(c1,c2,…,c 2i (7)

[0031] Where F represents fitness, g1 represents the total number of conflict points during the taxiing process of departing aircraft, g2 represents the number of delays of departing aircraft, and I represents the total number of departing aircraft; (c1,c2,…,c 2I (c1, c2, ..., c) represents an individual in the population, i.e., the scheduling scheme; where (c1, c2, ..., c) I (c) represents the starting taxiing time of each departing aircraft in the scheduling scheme. I+1 ,c I+2 ,…,c 2I ) represents the runway allocated to each departing aircraft in the scheduling scheme; c i With c I+i c is the value of the i-th position of this individual. i c is the starting taxiing time for departing aircraft i. I+i The runway assigned to departing aircraft i; i = 1, 2, ..., I; ω1 and ω2 represent the weights of the number of conflict points and the number of delayed aircraft, respectively;

[0032] Given the set of runway gates *e* and parking positions *s* for arriving aircraft, Dijkstra's algorithm is used to solve for the paths. These paths are then combined with the standard taxi paths for departing aircraft. The aforementioned conflict detection method is used to perform conflict avoidance, ultimately obtaining the number of conflict points and the waiting time for each aircraft. Then, the ground taxiing time of arriving aircraft is taken into account, and the arrival time of each departing aircraft at its assigned runway gate is calculated. The arrival time at the runway gate = c i +s ij / v+taxiing wait time; given the known arrival time of departing aircraft at the runway gate and the corresponding runway gate c I+i Without changing the landing time and runway gate of the departing aircraft, a first-come, first-served approach is adopted for scheduling. After scheduling, the queuing time of the departing aircraft can be obtained. Finally, the actual takeoff time of the departing aircraft is calculated as: arrival time at the runway gate + runway gate queuing time + t setup Then compare the actual takeoff time with T pi In comparison, when it is greater than T pi This describes aircraft takeoff delays and calculates the total number of departing aircraft delays.

[0033] The fitness of each individual is calculated, and a roulette wheel selection method is used on the entire population to select a certain number of new individuals to form a new population.

[0034] Furthermore, the method for determining conflict points is as follows:

[0035] The system acquires the nodes traversed by each arriving and departing aircraft during taxiing and obtains data on arriving aircraft. A set of potential conflict points is obtained by identifying duplicate points. When a potential conflict point exists in the set, it is determined whether it is a conflict point. The nodes traversed by departing aircraft are determined by the nodes traversed by all standard taxiing paths on the corresponding runway. The nodes traversed by arriving aircraft are determined by the nodes traversed by the shortest path between the landing runway and the parking stand. The determination method is as follows: First, using the landing time and taxiing path of the arriving aircraft, and the initial taxiing time and standard taxiing path of the departing aircraft, the arrival time of each arriving and departing aircraft passing through that point is calculated. The time difference is determined to be greater than the minimum safe interval. If it is less than the minimum safe interval, it is a conflict point. When a point is a conflict point, the arriving aircraft waits to avoid it, and the time interval between waiting and taxiing through the conflict point is not less than the safe interval.

[0036] Furthermore, the crossover operation of the genetic algorithm is as follows:

[0037] A single-point crossover method is used, where a point is randomly selected in the individual's coding string as the crossover point, and then the chromosome is split in two; the genetic information on the left side of the newly generated offspring chromosome comes from one parent chromosome, while the genetic information on the right side comes from another parent chromosome.

[0038] The chromosome coding method uses a two-layer coding structure. The first layer of coding is the starting taxi time of each aircraft, and the second layer of coding is the runway code assigned to each aircraft. The chromosome has a total of 2I bits, representing I departing aircraft. The first I digits of the chromosome represent the starting taxi time of the departing aircraft, and the last I digits represent the runway gate code of the aircraft.

[0039] Furthermore, the mutation operation of the genetic algorithm is as follows:

[0040] Single-point mutation refers to setting a random mutation point after confirming that a mutation operation is to be performed, and generating a random number within a specified range to replace the value at that mutation point.

[0041] Compared with the prior art, the present invention has the following technical features:

[0042] The main innovation of this invention is runway scheduling for aircraft at large airports. This invention can further reduce the number of conflict points and flight delays during aircraft taxiing, thereby improving airport safety and flight punctuality, and enabling airports to operate better. Attached Figure Description

[0043] Figure 1 To generate an initial solution flowchart;

[0044] Figure 2 Here is a flowchart of the genetic algorithm;

[0045] Figure 3 Chromosome coding diagram;

[0046] Figure 4 Diagram illustrating a skid collision;

[0047] Figure 5 This is a flowchart for calculating the fitness function. Detailed Implementation

[0048] This invention provides airports with a feasible runway selection scheme for departing aircraft, solving the runway scheduling problem. The relevant application background is that airports need to schedule both departing and arriving aircraft, and these aircraft can interfere with each other during taxiing. During taxiing, airports prioritize minimizing the number of conflict points and delays for departing aircraft. Therefore, this invention focuses on reducing the number of taxiing conflict points and delays for both arriving and departing aircraft. The problem addressed by this scheme has the following characteristics:

[0049] 1) Departing aircraft can delay their departure, but the landing time of arriving aircraft cannot be adjusted; 2) Departing aircraft have a defined departure taxiway and will taxi according to the standard taxiway by default; 3) Arriving aircraft can choose the fastest route back to their parking position after passing the runway center point; 4) The runway selection for both arriving and departing aircraft must strictly adhere to the wind direction of the day. Even if a northbound runway is available, a northbound runway cannot be selected in a southbound wind direction; 5) According to relevant literature, the wake turbulence safety separation is set at 96 seconds between large aircraft in front and small aircraft behind; The safe wake turbulence separation between aircraft is 196 seconds; the safe wake turbulence separation between smaller aircraft in front and larger aircraft behind is 60 seconds; the safe wake turbulence separation between smaller aircraft in front and smaller aircraft behind is 82 seconds, and these are used as the safe intervals; 6) If the actual takeoff time of an aircraft is longer than the estimated departure time, it is considered a delay; 7) Runway dispatchers only dispatch aircraft from their parking positions to the runway gate; 8) By default, the taxiing speed of aircraft is constant and the same; 9) By default, there are no other obstacles during the taxiing process of aircraft except for taxiing conflicts; 10) By default, the preparation time for takeoff of departing aircraft (flight straightening and other actions) is constant at 2 minutes.

[0050] Because runway scheduling is an extremely complex problem, finding an optimal solution is currently difficult; we can only strive to obtain a better, more suitable solution. Therefore, this invention combines a heuristic rule-based algorithm with a genetic algorithm. The better solution obtained by the heuristic rule-based algorithm is used as the initial solution to generate a parent population, and then the genetic algorithm is used for iterative solving. Experiments have shown that this method can effectively reduce conflicts between aircraft during taxiing, further ensuring aircraft safety. Furthermore, reasonable planning of departure flight sequences and initial taxiing times can reduce waiting time caused by runway congestion preventing timely taxiing, or waiting time during taxiing to avoid conflicts, thus reducing delays.

[0051] Referring to the accompanying drawings, the present invention provides a method for scheduling large-scale airport runways by fusing rule-triggered and genetic algorithms, comprising the following steps:

[0052] Step 1: Obtain the airport's standard taxiway path map and all data of airport arrivals and departures within the same time period, and perform preprocessing to remove irrelevant data.

[0053] In step 1, all relevant data for arriving and departing aircraft includes:

[0054] Departing aircraft data: parking position, assigned runway, aircraft type, aircraft departure time, and ground taxiing time; Arriving aircraft data: parking position, landing runway, and landing time.

[0055] For the preprocessed data, the aircraft type, takeoff time, and ground taxiing time of each aircraft in the departing aircraft data are grouped together and used as input for the heuristic rule algorithm in step 2. Similarly, the parking position, runway gate, and landing time of each aircraft in the arriving aircraft data are grouped together and used as one of the inputs for conflict resolution in step 4. The runway gate is used for takeoff during departure and for landing during arrival.

[0056] Each set of data mentioned above corresponds to one aircraft. That is, within the selected time period, for departing aircraft data, there are as many sets of data as there are aircraft; the same applies to arriving aircraft data.

[0057] Step 2: Use a heuristic rule algorithm to sort the priority sequence of each set of departing aircraft data obtained in Step 1, and obtain the priority ranking of each departing aircraft as shown below.

[0058] The safe time interval between departing aircraft is used as the setting time (used to calculate the average setting time), the average processing time is the average ground taxiing time of the original aircraft provided by the airport, and the expiration time is set to the departure time of the aircraft set by the airport. Since the relevant calculation formula needs to take into account the importance of each departing aircraft, i.e., the weight, to calculate the priority of each aircraft, and the airport has not given any special instructions on this, the weight of each aircraft is set to 1 in this scheme, and the aircraft priority is calculated using this parameter.

[0059] The heuristic rule algorithm obtains the priority ranking of departing aircraft using the following formula:

[0060]

[0061] The variables in the above formula are interpreted as follows:

[0062]

[0063]

[0064]

[0065] Among them, C is used in variables τ and R. max Variables, and because of C max It is difficult to estimate, so an estimated quantity is used. To express.

[0066] The priority index I for the next departing aircraft operation j, calculated using the above formula, is obtained after operation l is completed. j (t,l) sorts the indices from largest to smallest, which is the priority order of the corresponding departing aircraft.

[0067] Step 3: Prioritize each departing aircraft and allocate available runways to them according to their priority. The corresponding start taxiing time is set to 15 minutes before the aircraft's departure time. The allocation results constitute the scheduling plan, which includes the runway allocated to each departing aircraft and the start taxiing time. This plan serves as the initial scheme for subsequent algorithms. The corresponding process is as follows: Figure 1 As shown, the specific allocation algorithm is explained below:

[0068] First, a set of departing aircraft is created, containing relevant information for all departing aircraft to be sorted. Priority indices for each departing aircraft are obtained using a heuristic algorithm, and then sorted. The departing aircraft with the highest priority index is added to the planned aircraft set, then removed from the original departing aircraft set, and its current time is changed to the current time plus the aircraft's ground taxiing time. Priority indices are recalculated for the remaining departing aircraft in the departing aircraft set until all departing aircraft are placed into the planned aircraft set in priority order. Then, runway allocation for takeoff begins: first, a runway is allocated to the departing aircraft with the highest priority index; if a runway is available (i.e., no aircraft is ready to take off), then... The standard taxiway path is used to assign the aircraft to the nearest available runway. Then, the next highest priority departing aircraft is assigned a runway. If no runway is available, the allocation is delayed until the time interval between the aircraft on the current runway and the departing aircraft is greater than the safe time interval. This safe time interval is related to the size of the two aircraft. It is calculated by comparing the size of the aircraft on the current runway with the size of the departing aircraft to determine their set time. When the initial taxiway interval between the two aircraft is greater than the set time, it is assumed that the two aircraft will not interfere with each other. At this time, the runway corresponding to the departing aircraft can be considered an available runway for allocation. The above process is repeated until all departing aircraft are assigned their corresponding runways, thus completing the generation of the scheduling plan.

[0069] Step 4: Use the generated scheduling scheme as the initial solution and generate a certain number of populations. Put them into the genetic algorithm for iterative calculation. Finally, the optimal individual is obtained through the genetic algorithm, which is the final scheduling scheme.

[0070] Step 4.1, Initialize the population

[0071] A set of aircraft taxiing schemes is initialized based on the initial solution, which serves as the population for the genetic algorithm. The initial solution takes the form of {starting taxiing time 1, runway assigned to departing aircraft 1; starting taxiing time 2, runway assigned to departing aircraft 2; ... starting taxiing time n, runway assigned to departing aircraft n}. Considering that the starting taxiing time cannot be later than the takeoff time, the starting taxiing time is obtained by randomly selecting a value from 15 minutes before the aircraft's takeoff time to the aircraft's takeoff time. Due to airport restrictions, departing aircraft in each direction can only take off and depart from a fixed number of runways.

[0072] Therefore, the initialization of the population involves randomly changing the start taxiing time and assigned runway of any number of departing aircraft based on the original initial solution. After each change, a new scheduling scheme is generated. After a certain number of changes, all the new scheduling schemes and the initial solution together constitute the individuals of the initial population.

[0073] The method for each change is as follows:

[0074] The departure aircraft will be randomly selected at any time within 15 minutes before its departure time as its new starting taxi time; and one of the remaining runways in the same direction as the departure aircraft's assigned runway will be randomly selected as its new runway.

[0075] Step 4.2: Selection, mutation, crossover, and iteration.

[0076] Based on the initial population obtained in step 4.1, perform genetic operations such as selection, mutation, crossover, and iteration.

[0077] Step 4.2.1, select

[0078] The selection operation employs a roulette wheel selection method. The roulette wheel is divided into population sizes based on the proportion of individual fitness values. The size of the roulette wheel area is directly proportional to the individual's fitness value; that is, the higher the fitness of an individual, the larger the roulette wheel area it occupies, and thus the greater its probability of being selected. Before introducing fitness, conflict resolution must first be explained, as it plays a crucial role in the fitness calculation of this invention. Conflict resolution is explained below:

[0079] Conflicts on airport taxiways mainly manifest as conflicts between aircraft taxiing paths, including overlapping or intersecting paths. Figure 4 As shown.

[0080] Overtaking conflicts occur when a following aircraft has the potential to overtake an preceding aircraft while traversing the same taxiway. Head-on conflicts occur when two aircraft travel from opposite directions on the same taxiway; however, since airports use one-way taxiing, this type of conflict is not considered, and the calculations control each aircraft to taxi in the direction permitted by the taxiway. Crossing conflicts occur when the taxiways of two aircraft intersect at points.

[0081] If a conflict occurs, avoidance should be carried out in the following manner:

[0082] The system acquires the nodes traversed by each arriving and departing aircraft during taxiing and obtains data on arriving aircraft. A set of potential conflict points is obtained by identifying duplicate points. When a potential conflict point exists in the set, it is determined whether it is a conflict point. The nodes traversed by departing aircraft are determined by the nodes traversed by all standard taxiing paths on the corresponding runway. The nodes traversed by arriving aircraft are determined by the nodes traversed by the shortest path between the landing runway and the parking stand. The determination method is as follows: First, using the landing time and taxiing path of the arriving aircraft, and the initial taxiing time and standard taxiing path of the departing aircraft, the arrival time of each arriving and departing aircraft passing through that point is calculated. The time difference is determined to be greater than the minimum safe interval. If it is less than the minimum safe interval, it is a conflict point. When a point is a conflict point, the arriving aircraft waits to avoid it, and the time interval between waiting and taxiing through the conflict point is not less than the safe interval. If the waiting time exceeds a preset time threshold, causing the aircraft to take off later than its scheduled departure time, it is also considered a delay.

[0083] After identifying the conflict and its points, fitness can be calculated. The method for calculating fitness is as follows:

[0084] The fitness function can estimate the fitness of each individual in the population. The individual fitness function defined in this scheme is shown in equation (7):

[0085] F=ω1g1(c1,c2,…,c 2I )+ω2g2(c1,c2,…,c 2I (7)

[0086] Where F represents fitness, g1 represents the total number of conflict points during the taxiing process of departing aircraft, g2 represents the number of delays of departing aircraft, and I represents the total number of departing aircraft; (c1,c2,…,c 2I (c1, c2, ..., c) represents an individual in the population, i.e., the scheduling scheme; where (c1, c2, ..., c) I (c) represents the starting taxiing time of each departing aircraft in the scheduling scheme. I+1 ,c I+2 ,…,c 2I () represents the runway assigned to each departing aircraft in the scheduling scheme; that is, c i With c I+i c is the value of the i-th position of this individual. i c is the starting taxiing time for departing aircraft i. I+i The runway assigned to departing aircraft i; i = 1, 2, ... I; ω1 and ω2 represent the weights of the number of conflict points and the number of delayed aircraft, respectively.

[0087] Based on the known arrival and departure aircraft data, and considering the parking positions and assigned runways for departing aircraft, the standard taxiing path from departing aircraft i to runway j is r. ij and the corresponding distance s ij The scheduled departure time for the departing aircraft is T. pi The taxiing speed is v, and the runway preparation time for takeoff is t. setup The maximum safe wake separation is t wake The set of runways for arriving aircraft is e, the set of landing times is t, the set of parking positions is s, and the taxiing time of the arriving aircraft on the runway is t. land .

[0088] The process for calculating the fitness function is as follows: Figure 5 As shown, to consider the impact of arriving aircraft on departing aircraft during the ground taxiing phase, given the set of runway gates *e* and parking positions *s* for arriving aircraft, Dijkstra's algorithm is used to solve the path. This path is then merged with the standard taxiing path of departing aircraft. The aforementioned conflict detection method is used for conflict avoidance, ultimately obtaining the number of conflict points and the waiting time for each aircraft. Then, taking the ground taxiing time of arriving aircraft, the arrival time of each departing aircraft at its assigned runway gate is calculated, i.e., the arrival time at the runway gate = c. i +s ij / v+taxiing wait time; given the known arrival time of departing aircraft at the runway gate and the corresponding runway gate c I+i Without changing the landing time and runway threshold of inbound aircraft, departing aircraft are scheduled using a first-come, first-served approach. The scheduling process must consider the wake turbulence safety interval. After scheduling, the queuing time for departing aircraft can be obtained. Finally, the actual takeoff time of departing aircraft is calculated as: Arrival time at the runway threshold + Runway threshold queuing time + t setup Then compare the actual takeoff time with T pi In comparison, when it is greater than T pi To explain aircraft takeoff delays, calculate the total number of departing aircraft delays, and finally add the number of conflict points and the total flight delays together to obtain the fitness value.

[0089] Using the above methods, the fitness of each individual can be obtained. By applying the roulette wheel selection method to the entire population, a certain number of new individuals can be selected to form a new population.

[0090] Step 4.2.2, Cross

[0091] Crossover is a specific method that involves exchanging some of the genes (the initial taxiing time of departing aircraft, the assigned runway) between two individuals to create two completely new individuals.

[0092] This scheme employs a single-point crossover method, where a point is randomly selected in the individual's encoding string as the crossover point, subsequently splitting the chromosome in two. The left genetic information of the newly generated offspring chromosome comes from one parent chromosome, while the right genetic information comes from the other parent chromosome. The chromosome encoding information of this scheme is shown below:

[0093] Chromosome coding employs a two-layer coding structure. The first layer is the initial taxiing time for each aircraft, and the second layer is the runway code for each aircraft. Figure 3 As shown, the chromosome has 2I bits, representing one departing aircraft. The first I digits of the chromosome represent the aircraft's initial taxiing time, and the last I digits represent the aircraft's runway gate code.

[0094] Step 4.2.3, Mutation

[0095] Mutation is the process of creating a new individual after crossover, which has a certain probability of genetic mutation. This invention uses single-point mutation, which means that after confirming that a mutation operation is to be performed, a mutation point is randomly set, and a random number within a specified range is generated to replace the value at the mutation point.

[0096] The detailed flowchart for step 4 is as follows: Figure 2 As shown; except for the content specifically mentioned above, the genetic algorithm in this scheme is all existing technology and will not be described in detail.

[0097] Example:

[0098] The present invention compares the existing runway scheduling algorithm with the new runway scheduling algorithm used in the invention to analyze the conflict points, number of delayed flights, and taxiing time of two sets of standardized flight paths at the airport.

[0099] The number of conflict points, delayed flights, and taxiing time have all been significantly reduced, effectively meeting the airport's requirements for safety and reduced delays, while also improving airport efficiency and reducing taxiing time. Specific comparisons are shown below:

[0100] Table 1 Comparison of Conflict Points

[0101] Existing runway scheduling scheme This plan Road Network 1 0.296 0.259 Road Network 2 0.333 0.296

[0102] Table 2 Comparison of Delayed Flights

[0103] Existing runway scheduling scheme This plan Road Network 1 1.037 0.222 Road Network 2 0.963 0.296

[0104] Table 3 Comparison of Gliding Time

[0105]

[0106]

[0107] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for scheduling large-scale airport runways that integrates rule-based triggering and genetic algorithms, characterized in that, include: Obtain the airport's standard taxiway map and all data of airport arrivals and departures within the same time period, and preprocess them. The priority sequence of departing aircraft is sorted for each set of data using a heuristic rule algorithm to obtain the priority ranking of each departing aircraft. Prioritize each departing aircraft, allocate runways to departing aircraft according to the priority, and set the corresponding start taxiing time. The situation after allocation is used as the scheduling plan. The generated scheduling scheme is used as the initial solution. A population is generated based on the initial solution, and iterative calculations are performed through a genetic algorithm. Finally, the optimal individual is obtained through the genetic algorithm, which is the final scheduling scheme. The heuristic rule algorithm obtains the priority ranking of departing aircraft using the following formula: Among them, I j (t,l) represents the priority index of the next departing aircraft operation j after operation l is completed; j is the operation sequence number, indicating the process of the departing aircraft with the corresponding sequence number proceeding to the corresponding runway gate according to the standard taxiing path from the start taxiing time and then taking off for departure; l represents the index of the completed operation; w j p represents the weight of task j. j This represents the ground taxiing time of the departing aircraft corresponding to task j; d j Let t represent the departure time of the departing aircraft corresponding to task j, and t represent the current time. This indicates the average ground taxiing time for departing aircraft. s represents the average safe time interval between departing aircraft. lj This represents the safe time interval between job j and job l, where k1 and k2 are the look-ahead and scaling parameters, respectively; τ is the compactness factor, R is the range factor, η is the set-time severity factor, and n is the number of jobs. d represents the average takeoff time of departing aircraft, β represents the time correlation coefficient of variation, and d represents the average takeoff time of departing aircraft. max d min These represent the departure times of the latest and earliest departing aircraft, respectively. Priority index I j Sort (t,l) from largest to smallest, which is the priority sorting of the corresponding departing aircraft.

2. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm as described in claim 1, characterized in that, The preprocessing is an irrelevant data removal operation; The preprocessed data includes: Departing aircraft data: parking position, allocated runway, aircraft type, aircraft departure time, and ground taxiing time; Arriving aircraft data: parking position, landing runway, landing time.

3. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm as described in claim 1, characterized in that, The aircraft type, takeoff time, and ground taxiing time of each aircraft in the departing aircraft data are used as a set of data as input to the heuristic rule algorithm; the parking position, runway gate, and landing time of each aircraft in the arriving aircraft data are used as a set of data as input to the conflict judgment.

4. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm as described in claim 1, characterized in that, The allocation of runways to departing aircraft according to their order includes: First, a set of departing aircraft is created, containing relevant information for all departing aircraft to be sorted. Priority indices for each departing aircraft are obtained using a heuristic algorithm, and then sorted. The departing aircraft with the highest priority index is added to the planned aircraft set, then removed from the original departing aircraft set, and its current time is changed to the current time plus the aircraft's ground taxiing time. The priority indices of the remaining departing aircraft in the departing aircraft set are recalculated until all departing aircraft are placed into the planned aircraft set in priority order. Then, runway allocation begins: first, a runway is allocated to the departing aircraft with the highest priority index. If a runway is available (i.e., no aircraft is ready to take off), it is allocated to the nearest available runway according to the standard taxiway path diagram. Then, runway allocation is performed for the next highest priority departing aircraft. If no runway is available, allocation is performed only after the time interval between the aircraft and the aircraft on a non-available runway exceeds the safe time interval. This process is repeated until all departing aircraft are allocated their corresponding runways, completing the generation of the scheduling plan.

5. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm according to claim 1, characterized in that, The initialization population process of the genetic algorithm includes: A set of aircraft taxiing schemes is initialized based on the initial solution, which serves as the population for the genetic algorithm. The initial solution takes the form of {starting taxiing time 1, runway assigned to departing aircraft 1; starting taxiing time 2, runway assigned to departing aircraft 2; ... starting taxiing time n, runway assigned to departing aircraft n}. Based on the initial solution, the starting taxiing time and assigned runway of any number of departing aircraft are randomly changed. A new scheduling scheme is generated after each change. After a certain number of changes, all the new scheduling schemes and the initial solution are used as individuals in the initial population. The method for each change is as follows: The departure aircraft will be randomly selected at any time within 15 minutes before its departure time as its new starting taxi time; and one of the remaining runways in the same direction as the departure aircraft's assigned runway will be randomly selected as its new runway.

6. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm according to claim 1, characterized in that, The selection operation of the genetic algorithm includes: The selection operation uses the roulette wheel selection method. The roulette wheel is divided into population sizes based on the proportion of individual fitness values, and the size of the roulette wheel area is directly proportional to the fitness of an individual. The fitness is calculated as follows: F=ω1g1(c1,c2,…,c 2I )+ω2g2(c1,c2,…,c 2I ) (7) Where F represents fitness, g1 represents the total number of conflict points during the taxiing process of departing aircraft, g2 represents the number of delays of departing aircraft, and I represents the total number of departing aircraft; (c1,c2,…,c 2I (c1, c2, ..., c) represents an individual in the population, i.e., the scheduling scheme; where (c1, c2, ..., c) I (c) represents the starting taxiing time of each departing aircraft in the scheduling scheme. I+1 ,c I+2 ,…,c 2I ) represents the runway allocated to each departing aircraft in the scheduling scheme; c i With c I+i c is the value of the i-th position of this individual. i c is the starting taxiing time for departing aircraft i. I+i The runway assigned to departing aircraft i; i = 1, 2, ... I; ω1 and ω2 represent the weights of the number of conflict points and the number of delayed aircraft, respectively.

7. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm as described in claim 6, characterized in that, The method for determining conflict points is as follows: The system acquires the nodes traversed by each arriving and departing aircraft during taxiing and obtains data on arriving aircraft. A set of potential conflict points is obtained by identifying duplicate points. When a potential conflict point exists in the set, it is determined whether it is a conflict point. The nodes traversed by departing aircraft are determined by the nodes traversed by all standard taxiing paths on the corresponding runway. The nodes traversed by arriving aircraft are determined by the nodes traversed by the shortest path between the landing runway and the parking stand. The determination method is as follows: First, using the landing time and taxiing path of the arriving aircraft, and the initial taxiing time and standard taxiing path of the departing aircraft, the arrival time of each arriving and departing aircraft passing through that point is calculated. The time difference is determined to be greater than the minimum safe interval. If it is less than the minimum safe interval, it is a conflict point. When a point is a conflict point, the arriving aircraft waits to avoid it, and the time interval between waiting and taxiing through the conflict point is not less than the safe interval. Given the set of runway gates *e* and parking positions *s* for arriving aircraft, Dijkstra's algorithm is used to solve for the paths. These paths are then combined with the standard taxi paths for departing aircraft. The aforementioned conflict detection method is used to perform conflict avoidance, ultimately obtaining the number of conflict points and the waiting time for each aircraft. Then, the ground taxiing time of arriving aircraft is taken into account, and the arrival time of each departing aircraft at its assigned runway gate is calculated. The arrival time at the runway gate = c i +s ij / v+taxiing wait time; given the known arrival time of departing aircraft at the runway gate and the corresponding runway gate c I+i Without changing the landing time and runway gate of the departing aircraft, a first-come, first-served approach is adopted for scheduling. After scheduling, the queuing time of the departing aircraft can be obtained. Finally, the actual takeoff time of the departing aircraft is calculated as: arrival time at the runway gate + runway gate queuing time + t setup Then compare the actual takeoff time with T pi In comparison, when it is greater than T pi This describes aircraft takeoff delays and calculates the total number of departing aircraft delays. The fitness of each individual is calculated, and a roulette wheel selection method is used on the entire population to select a certain number of new individuals to form a new population.

8. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm according to claim 1, characterized in that, The crossover operation of the genetic algorithm is as follows: A single-point crossover method is used, where a point is randomly selected in the individual's coding string as the crossover point, and then the chromosome is split in two; the genetic information on the left side of the newly generated offspring chromosome comes from one parent chromosome, while the genetic information on the right side comes from another parent chromosome. The chromosome coding method uses a two-layer coding structure. The first layer of coding is the starting taxi time of each aircraft, and the second layer of coding is the runway code assigned to each aircraft. The chromosome has a total of 2I bits, representing I departing aircraft. The first I digits of the chromosome represent the starting taxi time of the departing aircraft, and the last I digits represent the runway gate code of the aircraft.

9. The large-scale airport runway scheduling method based on fusion rule triggering and genetic algorithm according to claim 1, characterized in that, The mutation operation of the genetic algorithm is as follows: Single-point mutation refers to setting a random mutation point after confirming that a mutation operation is to be performed, and generating a random number within a specified range to replace the value at that mutation point.

Images