A real-time parking space allocation method for airport
By constructing a real-time parking space allocation model with the goal of minimizing aircraft taxiing distance and maximizing parking space occupancy, and using a genetic algorithm for optimization calculation, the problem of low parking space allocation efficiency in existing technologies is solved. This achieves real-time airport allocation with the shortest taxiing distance and the highest parking space utilization efficiency, demonstrating practicality and economy.
Patent Information
- Application Number
- CN202210722776.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-24
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-06-24
AI Technical Summary
Existing research on airport parking space allocation mostly focuses on pre-allocation, with limited research on real-time allocation. Furthermore, existing methods fail to effectively utilize heuristic algorithms for optimization, neglecting aircraft taxiing distance and parking space occupancy rates, resulting in low allocation efficiency in emergency situations.
A real-time allocation model is constructed with the optimization objectives of minimizing aircraft taxiing distance and maximizing parking space occupancy rate. The model is optimized using a genetic algorithm. Combining the objective functions of taxiing distance and parking space utilization efficiency with safety constraints, the optimal parking space allocation result is obtained through gene encoding, initial population generation, fitness function, selection, crossover, and mutation calculations.
It enables efficient real-time allocation of parking positions in case of emergencies, reduces taxiing distance and parking position occupancy, improves airport operational efficiency and safety, and saves taxiing fuel costs.
Smart Images

Figure CN114943388B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of airports, and in particular relates to a real-time parking space allocation method for an airport. Background Art
[0002] In recent years, more and more scholars at home and abroad have studied the parking space problem by modeling and using different optimization algorithms. Although the research on parking space allocation at home and abroad has been increasing in recent years, there are still some shortcomings:
[0003] At present, the research on parking space allocation is mostly at the level of pre-allocation, and there is less research on real-time parking space allocation. In the case of flight delays due to traffic control, weather changes or emergencies, the results of pre-allocation often fail and parking space conflicts occur. At present, most airports in my country adjust parking spaces manually in real time, and the scientificity and rationality need to be improved.
[0004] In terms of the algorithms selected for research, because the exact algorithm appeared relatively early, many scholars still use the exact algorithm to optimize the calculation of the parking space allocation problem. Considering the construction of smart civil aviation in my country and the continuous development of artificial intelligence, we should seek more heuristic algorithms to optimize the analysis of the parking space problem, such as genetic algorithms, simulated annealing algorithms, ant colony algorithms and other heuristic algorithms.
[0005] Although existing studies on parking space allocation have different research objectives, mainly for different stakeholders such as passengers and airports, most scholars choose to conduct research based on the shortest walking distance for passengers. In the past, walking distance was chosen to improve the comfort of passengers boarding the aircraft. However, with the popularization of convenient facilities such as horizontal elevators in airports, it is more convenient for passengers to transfer or board the aircraft. Therefore, research should no longer focus on the passenger boarding distance. Summary of the invention
[0006] In view of this, the present invention aims to propose a real-time airport parking space allocation method to solve the problems mentioned in the above background technology.
[0007] To achieve the above object, the technical solution of the present invention is achieved as follows:
[0008] A real-time airport parking space allocation method, characterized by: comprising:
[0009] A real-time parking stand allocation model is constructed with the shortest aircraft taxiing distance and the highest parking stand occupancy rate as the optimization objectives and the safety of surface operations as the constraint condition.
[0010] The parking stand allocation model is optimized using genetic algorithm to obtain the parking stand pre-allocation result.
[0011] Furthermore, the function corresponding to the optimization objective is
[0012] f=f1+f2 (1)
[0013] f1=min[∑ i∈N ∑ k∈M x ik (DA k +DD K )] (2)
[0014]
[0015] Where N represents the flight set (i, j∈N; i=1,2,3…20; j=2,3,4…20)
[0016] M represents the parking space set (k∈M; k=1,2,3…28)
[0017] DA represents the flight sliding distance set (DA k ∈DA; k∈M; k=1,2,3…28)
[0018] DD represents the flight taxiing distance set (DD k ∈DD; k∈M; k=1,2,3…28)
[0019]
[0020] Among them, f1 represents the optimization goal of the shortest taxiing distance, and the objective function is to find the minimum sum of the taxiing distance and the taxiing distance of each flight, that is, the shortest total taxiing distance;
[0021] f2 means maximizing the efficiency of parking space utilization and reducing parking space occupancy.
[0022] Furthermore, the constraint condition is specifically:
[0023]
[0024]
[0025]
[0026]
[0027]
[0028]
[0029] in,
[0030]
[0031]
[0032] E ik represents the off-block time for flight i to enter parking stand k;
[0033] L ik represents the off-block time of flight i leaving parking stand k;
[0034] E jk represents the block time of flight j that uses the same parking stand k after flight i;
[0035] Among them, formula (4) constrains that a flight can only be arranged at one parking position; formula (5) means that a parking position can only accommodate one flight at most in the same time period; formula (6) means that there needs to be a 35-minute safety time interval between two adjacent flights; formula (7) means that each flight must stay at the parking position for at least one hour, which is the constraint of the minimum stopover time; formulas (8) and (9) mean that the airport's operating hours are from 5:00 to 24:00 every day, and flights are only allowed to take off and land within this time range.
[0036] Furthermore, the genetic algorithm is used for calculation in step 3, which specifically includes the following:
[0037] Step 1: Gene encoding using the taxiing distance of each parking position;
[0038] Step 2: Generate the initial population randomly;
[0039] Step 3: Set the individual fitness function in the population;
[0040] Step 4: Obtain the optimal solution through selection, crossover, and mutation calculation, which is the final result of parking space pre-allocation.
[0041] The present invention also provides a real-time parking space allocation device for an airport, comprising:
[0042] A model building device is used to construct a real-time parking stand allocation model with the shortest aircraft taxiing distance and the highest parking stand occupancy rate as optimization objectives and the safety of surface operations as a constraint condition;
[0043] The model solving device is used to optimize the parking space allocation model by using a genetic algorithm to obtain a parking space pre-allocation result.
[0044] The present invention also provides an electronic device, comprising
[0045] at least one processor, and
[0046] at least one memory in communication with the processor, wherein:
[0047] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute any of the methods described above.
[0048] The present invention also provides a non-volatile computer-readable storage medium, which, when the computer-executable instructions are executed by one or more processors, enables the processors to execute any of the methods described above.
[0049] Compared with the prior art, the real-time airport parking space allocation method of the present invention has the following advantages:
[0050] (1) The present invention constructs an objective function with the shortest aircraft taxiing distance and the highest parking stand occupancy rate as the optimization objectives, and at the same time considers the safety constraints of the surface operation to establish a real-time parking stand allocation model. Therefore, the real-time allocation model is based on the premise of surface safety operation and ensures the smooth progress of production;
[0051] (2) The real-time parking stand allocation model constructed by the present invention can not only pre-allocate flights on the same day at the airport, but also allocate flights in real time in the case of small delays. It is a parking stand allocation model with real-time allocation function. It has certain effects in reducing aircraft congestion at airports and reducing the taxiing fuel costs of airlines, and provides a reference for the study of real-time parking stand allocation problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:
[0053] Figure 1 A flow chart for the generation of the initial population of the present invention;
[0054] Figure 2 It is a schematic diagram of the roulette selection method of the present invention;
[0055] Figure 3 It is a schematic diagram of cross calculation of the present invention;
[0056] Figure 4 It is a schematic diagram of the variation calculation process of the present invention;
[0057] Figure 5 It is the parking space diagram ZBTJ-2 of Tianjin Binhai Airport of the present invention;
[0058] Figure 6 This is a schematic diagram of measuring the first sliding distance of the present invention;
[0059] Figure 7 is a schematic diagram of a second sliding distance measurement of the present invention;
[0060] Figure 8 The glide path diagram of the present invention
[0061] Fig. 9 It is an iterative variation diagram of the present invention;
[0062] Fig.10 The Gantt chart of the parking position before optimization of the present invention;
[0063] Fig.11 The Gantt chart of the optimized parking spaces (pre-allocated parking spaces) of the present invention;
[0064] Fig.12 A graph is generated for the delay of the present invention. DETAILED DESCRIPTION
[0065] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.
[0066] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.
[0067] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.
[0068] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.
[0069] The present invention provides a method for real-time parking space allocation at an airport. First, the symbols used in the modeling are explained. Based on the data used in the present invention, the following variables are defined:
[0070] (1) N represents the flight set (i, j∈N; i=1,2,3…20; j=2,3,4…20)
[0071] (2) M represents the parking space set (k∈M; k=1,2,3…28)
[0072] (3)DA represents the flight sliding distance set (DA k ∈DA; k∈M; k=1,2,3… 28)
[0073] (4)DD represents the flight taxiing distance set (DD k ∈DD; k∈M; k=1,2,3… 28)
[0074]
[0075]
[0076]
[0077] (7)E ik represents the block time of flight i entering parking stand k
[0078] (8)L ik represents the off-block time of flight i leaving parking stand k
[0079] (9)E jk The block time of flight j entering the same parking stand k after flight i
[0080] The function corresponding to the optimization target of the present invention is:
[0081] f=f1+f2 (1)
[0082] f1=min[∑ i∈N ∑ k∈M x ik (DA k +DD K )] (2)
[0083]
[0084] Among them, f1 represents the optimization goal of the shortest taxiing distance, and the objective function is to find the minimum sum of the taxiing distance and the taxiing distance of each flight, that is, the shortest total taxiing distance. On the one hand, the shortest taxiing distance means the least taxiing fuel, which can save costs for airlines; on the other hand, it also means the shortest taxiing time, which can reduce the number of aircraft on the airport runway in a certain period of time, increase the runway capacity, speed up the speed of aircraft entering and leaving the parking space, and improve the efficiency of the use of the parking space.
[0085] f2 means maximizing the efficiency of parking space utilization and reducing parking space occupancy. Because each parking space corresponds to a different flight taxiing distance, under the goal of minimizing the total taxiing distance, the selection of parking spaces will be based on satisfying the constraints, and will be more inclined to certain parking spaces with relatively shorter taxiing distances. This results in a reduction in the number of parking spaces used and the occupancy of parking spaces.
[0086] The constraints used in the present invention are:
[0087]
[0088]
[0089]
[0090]
[0091]
[0092]
[0093] Among them, formula (4) constrains that a flight can only be arranged at one parking position; formula (5) means that a parking position can only accommodate one flight at most in the same time period; formula (6) means that there needs to be a 35-minute safety time interval between two adjacent flights; formula (7) means that each flight must stay at the parking position for at least one hour, which is the constraint of the minimum stopover time; formulas (8) and (9) mean that the airport's operating hours are from 5:00 to 24:00 every day, and flights are only allowed to take off and land within this time range.
[0094] The present invention selects a genetic algorithm to optimize the parking space allocation model. The advantage of the genetic algorithm is that due to its own evolutionary characteristics, it does not need to consider the intrinsic properties of the problem during calculation, and any form of objective function can be calculated. The concept of the genetic algorithm comes from Darwin's theory of evolution. According to the detailed problem that needs to be studied by the genetic algorithm, the data in the problem can be encoded, and the encoding can be regarded as a gene, and the result of the encoding can be regarded as an individual, and a group of encodings is a different population. Based on the population, the calculation is performed based on the idea of survival of the fittest in the theory of evolution. Through artificial function settings, it acts as the law of nature. After multiple generations of inheritance, good results are selected from a variety of possible results, and bad results are eliminated, so as to obtain the optimal solution.
[0095] The Python implementation process of the present invention using genetic algorithm is as follows
[0096] (1) Gene coding
[0097] Genetic coding is the foundation of genetic algorithms. To use genetic algorithms, coding must be performed first, because the coding situation determines the storage and expression of genes in the algorithm, which directly affects the subsequent overall calculation process. Genetic algorithm gene coding has many forms, which can be encoded in binary or real numbers. If binary coding is used, due to its large randomness, it will affect the crossover and mutation process. When the calculation is completed, it may be far away from the optimal solution and cannot reach stability. Therefore, the present invention adopts a self-created real number coding method, directly using the taxiing distance of each parking position for encoding, that is, the taxiing distance data of each parking position is the code of the parking position, which is convenient for subsequent data selection, crossover, mutation and other calculations, and is more convenient to obtain optimization results, avoiding the situation of no feasible solution, as shown in Table 1.
[0098] For example, the present invention allocates parking spaces for 20 flights and minimizes the taxiing distance as an example to illustrate the algorithm. Because the present invention only focuses on the taxiing distance, the meaning of chromosomes is relatively simple, so one chromosome represents an individual, and the genetic composition of each individual is the total taxiing distance corresponding to each of the 20 parking spaces.
[0099] Table 1 Gene coding table (partial)
[0100]
[0101] (2) Generation of the initial population
[0102] Because the genetic algorithm is inspired by the reproduction of nature, the first step in the computational solution is to determine an initial population as the beginning of reproduction. The initial population of this study was generated in a completely random manner. Some scholars will use some constraints to select relatively excellent genes when generating the initial population. Taking this study as an example, some aircraft positions with a short taxiing distance are first selected to join the initial population. In this way, the diversity of the population will be reduced, and premature maturity may occur. The above situation can be avoided by using a completely random method. The process of generating the initial population is as follows: Figure 1 shown.
[0103] The generation process is explained. In the example operation process, the initial population is based on the flight data. Therefore, after allocating parking spaces for 20 flights in the first step, the constraints in the second step are required to meet the basic constraints. And the constraints here do not include constraints on the parking space occupation time, that is, the initial parking space usage time is allowed to overlap. This overlap occurs because although the constraints prevent two flights from entering the same parking space at the same time, due to the different stopover times of each flight, there may be overlaps in the parking space occupation time. However, since the initial population is only a prerequisite for starting the operation, the overlapping part participates in the calculation and does not affect the subsequent calculation results.
[0104] After confirming that the parking space meets the requirements, execute the third step and repeat the first and second steps 50 times to obtain an initial population of 50 individuals. Subsequent calculations and iterations are carried out around this initial population.
[0105] (3) Fitness function
[0106] In the genetic algorithm, how to evaluate the quality of an individual depends entirely on the fitness function. The size of the fitness function represents the ability of the individual to adapt to the environment and determines the survival or extinction of the individual. The greater the fitness, the easier it is to survive through natural selection, and vice versa. In order to obtain the final result and speed up the solution, it is necessary to set the fitness function reasonably. Because the optimization goal of the present invention is to find the minimum value of the sliding distance, the fitness function is also to find the minimum value, that is, the shorter the total sliding distance of the individual, the stronger the adaptability, and the easier it is to survive in the process of natural selection, which is reflected in the code as the greater the fitness.
[0107] (4) Select calculation
[0108] Just as the process of survival of the fittest in nature is screened through the natural environment, the role of selection calculation in genetic algorithms is to simulate the process of natural selection, and the scale of the selection process is the fitness function. If an individual has a larger fitness, he is more likely to pass on his genes to the next generation through natural selection.
[0109] However, this kind of selection in nature is probabilistic. In order to realize the selection process in the algorithm, it is necessary to establish this probabilistic relationship. In this study, the selection calculation method adopts Roulette Wheel Selection, which is a playback random sampling method with relatively simple operation. The probability of each individual entering the next generation is equal to the ratio of its fitness value to the sum of the fitness values of individuals in the entire population. For example, there are five individuals that need to be selected and calculated to determine whether they can pass on their genes to the next generation. After the fitness function is calculated, the fitness of these five individuals is p1, p2...p5, then the probability of each individual being selected is:
[0110]
[0111] The roulette wheel composed of the probabilities of the five individuals being selected is as follows Figure 2 As shown. It can be seen that the greater the fitness of an individual, the greater the advantage it has among the individuals to be selected, and the easier it is to be selected in the natural selection process.
[0112] (5) Cross calculation
[0113] Crossover calculation refers to exchanging some genes of two paired individuals of the previous generation in a certain way to form a new individual. In the genetic algorithm, crossover calculation is the most important part of the algorithm, which gives the genetic algorithm the ability to search globally. Whether it is the generation of offspring individuals or mutation calculation, it needs to be performed after gene crossover. There are many ways to crossover. This study chooses single-point crossover as the crossover method, that is, randomly selects 1 gene from the 20 genes of the previous generation to participate in the crossover calculation. The crossover probability of the present invention is set to 70%, and the specific crossover operation is shown in the figure.
[0114] (6) Mutation calculation
[0115] Mutation calculation is to change the coding of some offspring individuals with a certain probability to achieve the effect of gene mutation. Mutation calculation has two functions. First, just like the mutation of organisms in nature during reproduction, mutation can prevent the early appearance of "super individuals" in the population, increase the diversity of the population, and prevent the algorithm from being premature. Second, it adds local search capabilities to the genetic algorithm: when the previous generation has undergone crossover calculation, the offspring may be very close to the optimal solution. At this time, changing the genes of certain positions of the offspring through mutation can accelerate the speed at which the offspring obtain the optimal solution.
[0116] Generally, mutation calculation is divided into two steps. First, whether the offspring individuals generated in the crossover process of the previous generation mutate depends on the mutation probability. Second, if mutation occurs, the coding of the offspring is changed according to the mutation form set in the algorithm. In the present invention, the mutation probability is set to 3%. Individuals that meet the mutation probability will have a gene mutated into an allele. The process of mutation calculation is as follows: Figure 4 shown.
[0117] The parameter settings of the genetic algorithm are shown in Table 2. The reason why the number of iterations is selected as 800 generations is that it is determined after analyzing the iteration results after multiple trial calculations. Considering the possible premature situation of the genetic algorithm during calculation, in order to make the result as close to the optimal solution as possible, the number of iterations is set to 800 generations.
[0118] Table 2 Genetic algorithm parameter settings
[0119]
[0120] The following application analysis is carried out by taking Tianjin Binhai International Airport as an example. The requirements for operating conditions at each airport are not exactly the same. When applied to other airports, the operating conditions can be revised in light of the airport's conditions.
[0121] (1) The flight take-off and landing runway has been determined by the control department. The present invention uses the 16R runway of Tianjin Binhai Airport for take-off and landing.
[0122] (2) The flight schedule, aircraft type, and parking space information are known. This study sets up 20 flights and 28 parking spaces. It is assumed that all flight information, including schedule, aircraft type, etc., is known; it is assumed that all flights in this study are A320 series or B737 series aircraft; it is assumed that the number and location of available parking spaces are also known. In this study, a randomly generated flight sequence is used, but actual flight information can also be imported.
[0123] (3) The flight maintains a constant speed during taxiing. It is assumed that the fuel consumption in this study is the average fuel consumption, and the difference in aircraft fuel consumption caused by turning, acceleration, deceleration, etc. is not considered. The taxiing fuel is only related to the taxiing distance.
[0124] After determining the assumptions, the constraints are explained. The constraints determined in this study are as follows:
[0125] (1) A flight can only stop at one parking stand.
[0126] (2) A parking stand can only be occupied by one flight at a time.
[0127] (3) If the same parking stand is used by two flights consecutively, there will be a 35-minute safety interval between the two flights.
[0128] (4) Tianjin Binhai Airport is open from 5:00 to 24:00 every day.
[0129] (5) The time the flight occupies the parking space is greater than the minimum stopover time (60 minutes).
[0130] (6) Each flight represents a flight pair, which includes two tasks: an incoming flight and a departing flight. The time a parking space is occupied is from the time the flight arrives to the time the flight departs.
[0131] (7) The present invention uses Tianjin Binhai Airport as the experimental object and assumes that all flights take off and land on runway 16R. Because taxiing distance is selected as the main research target during modeling, before using the genetic algorithm for optimization calculation, it is necessary to first obtain the taxiing distance data corresponding to different parking positions when flights take off and land on runway 16R.
[0132] Based on the parking space map of Tianjin Binhai Airport ZBTJ-2 in EAIP (2020-10.v1.18 version) Figure 5 Using the satellite map in Amap as a template, use the distance measurement tool provided by Amap to measure the flight taxiing distance in proportion ( Figure 6 , Figure 7 ) to obtain the glide distance data.
[0133] The present invention selects 28 parking stands in total, all of which are located in the T2 terminal of Tianjin Binhai Airport. Among them, since the fixed 16R runway is selected for take-off and landing, parking stands 201 to 218 are closer to the runway; and parking stands 219 to 228 are farther from the runway.
[0134] When measuring the taxiing distance, for incoming flights, the starting point of measurement is the end of the touchdown zone of runway 16R, and the end point is the jet bridge of the parking stand; for departing flights, the starting point of measurement is the jet bridge of the parking stand, and the end point is the head of runway 16R. When selecting the taxiing path, if the incoming flight stops at stands 201-211, the taxiing path consists of B4-C2-T3; if it stops at stands 212-219, the taxiing path consists of B4-C2-N3-Q; if it stops at stands 220-228, the taxiing path consists of B4-C2-N3-Q-N5-T5. The taxiing path for departing flights is selected in the same direction as that for arriving flights. In order to reduce taxiing conflicts, the taxiing path for arriving and departing flights is avoided as much as possible during selection. Generally, another path next to the taxiing path of the arriving flight is selected. The reference for selecting the taxiing path diagram is as follows: Figure 8 shown.
[0135] After all the taxiing distances of the parking stands are measured, the total taxiing distance data are summarized, as shown in Table 3. At present, the parking stand allocation model has been established and the key taxiing distance data has been obtained. The next section will start the implementation of the genetic algorithm optimization calculation process.
[0136] Table 3 Summary of sliding distance
[0137]
[0138]
[0139] The parameter settings of the genetic algorithm are shown in Table 2.
[0140] The present invention simulates the parking stand pre-allocation of Tianjin Binhai Airport, assuming that parking stands 201 to 228 are available for aircraft to park, and randomly generates 20 flight pairs running between 5:00 and 24:00. The flight information randomly generated by the case is shown in Table 4, and the initial parking stand information randomly allocated to these flights is shown in Table 5.
[0141] Table 4 Flight information table
[0142]
[0143] Table 4 (Continued) Flight Information Table
[0144]
[0145]
[0146] Table 5 Initial parking space table
[0147]
[0148] Pre-allocation and result analysis:
[0149] According to the parameter settings in Table 2, based on the flight information in Table 4, the genetic algorithm calculation is performed, and the results are obtained after running the Python program.
[0150] First, the iteration of the algorithm is analyzed. The iteration changes of the genetic algorithm are as follows: Fig. 9 shown.
[0151] Fig. 9 The vertical axis is the sliding distance, ranging from 108 km to 124 km, increasing by 2 km; the horizontal axis is the number of iterations, ranging from 0 to 800 generations, increasing by 100 generations.
[0152] As can be seen from the figure, the algorithm tends to be flat and has basically converged at around 350 generations. When the set 800 iterations are completed, the change in sliding distance can be ignored. It can be considered that the algorithm has converged and obtained an approximate optimal solution.
[0153] The approximate optimal solution obtained by genetic algorithm simulation calculation is the optimized parking position information, and the optimization result is the final result of parking position pre-allocation, as shown in Table 6.
[0154] Table 6 Parking space pre-allocation results
[0155]
[0156]
[0157] In order to intuitively display the optimization calculation process and results of the genetic algorithm, the data in Tables 4, 5 and 6 are plotted into Gantt charts of the parking spaces before and after optimization to compare the parking space allocation before and after optimization, as shown in Figure 2. Fig.10 , 11 shown.
[0158] First, by comparing the two figures above, it can be seen that before optimization, the flights only meet the constraints, including exclusivity constraints, safety interval constraints, etc. The flights are distributed on 14 parking stands. Although the constraints are met, the flights are relatively scattered and occupy many parking stands. After optimization according to the goal of the shortest taxiing distance, the flights are only distributed on 8 parking stands on the basis of meeting the constraints. This shows that the waste of parking stand resources is reduced, the utilization efficiency of the parking stands is increased, and the number of flights carried by each parking stand is also relatively uniform, which optimizes the configuration of airport parking stand resources and meets the optimization goal of reducing parking stand occupancy.
[0159] Secondly, the optimization goal of this study is also based on the perspective of airports and airlines, considering the allocation of parking spaces to minimize the taxiing distance of flights entering and leaving the airport (equivalent to the shortest taxiing time and the least taxiing fuel), reducing the possibility of aircraft congestion on the runway, and saving the taxiing fuel cost generated by taxiing. In order to verify the effectiveness of the genetic algorithm and whether the pre-allocation results are reasonable, it is necessary to compare the taxiing distance data before and after optimization. The data obtained after the algorithm calculation is shown in Table 7:
[0160] Table 7 Pre-allocation optimization results
[0161]
[0162] It can be seen from the table that the taxiing distance corresponding to the randomly assigned initial parking position before optimization is 122,472 meters; the taxiing distance after optimization calculated by the genetic algorithm is 108,268 meters. The total optimized taxiing distance is 14,204 meters, which is about 14 kilometers. It can be seen from the taxiing distance alone that the optimization effect is very obvious.
[0163] Referring to the hypothetical data in the flight planning course of Civil Aviation University of China, assuming that the aircraft taxiing speed is constant at 18 kilometers per hour, the taxiing distance is converted into more intuitive taxiing time and taxiing fuel. It can be concluded that this example saves about 47 minutes of taxiing time and about 852 kilograms of taxiing fuel. This example only simulates the operation of 20 flight pairs at Tianjin Binhai Airport. In October 2021, the weekly flight volume of Tianjin Binhai Airport will reach 3,000 flights. Based on the huge number of flights, the taxiing fuel that can be saved will be more obvious. In the context of the current rise in international oil prices, it will save airlines more taxiing fuel costs.
[0164] The present invention takes the taxiing distance as the optimization target and uses the genetic algorithm to perform optimization calculation. The obtained optimization results prove that the taxiing fuel consumption can be reduced by pre-allocating the parking positions, which has certain practical significance.
[0165] Real-time allocation and result analysis
[0166] In order to simulate flight delays, it is first necessary to generate delay conditions based on the original flight information. Based on the flight information used for pre-allocation of parking spaces, 4 flights are randomly selected from 20 flights, each with a delay of 10 to 30 minutes. As for how to define flight delays, the "Civil Aviation Flight Normal Statistics Method" issued by the Civil Aviation Administration of my country defines a normal flight as a flight that starts within the specified airport ground taxiing time after the planned door closing time, and does not encounter abnormal situations such as return or alternate landing, or lands no later than 10 minutes after the planned door opening time.
[0167] Although there is a definition of normal flights, the Civil Aviation Administration of my country does not specify in detail in the document how long the difference between the actual blocking / off-blocking time and the planned blocking / off-blocking time of a flight will be considered a flight delay. Regarding the determination of flight delays, the current civil aviation industry in my country generally believes, based on actual work conditions, that if the actual time of a flight differs from the planned time by more than 30 minutes, it is considered a flight delay. Therefore, in the present invention, the time of flight delay is deemed to be added on the basis of 30 minutes exceeding the original planned time of the flight. That is to say, the present invention generates a delay of 10 to 30 minutes, which is actually a delay of 40 to 60 minutes for the planned off-blocking time and planned on-block time of the flight. Use Python to write code, and after randomly generating delays, the actual flight information is as follows: Fig.12 shown.
[0168] The generated flight delay information was sorted out, among which: flight 3 was delayed by 16 minutes; flight 7 was delayed by 14 minutes; flight 11 was delayed by 21 minutes; flight 17 was delayed by 25 minutes. The detailed information of the delayed flights was summarized, and the flight delay situation is shown in Table 8.
[0169] Table 8 Flight Delays
[0170]
[0171] At this point, the actual flight information after the flight delay is obtained. Taking the delayed flight information as an example, the parking space allocation model proposed in this study is used, and the basic genetic algorithm written in Python is used to calculate the example. The genetic algorithm parameter settings are shown in Table 2. At the same time, since the delay situation is randomly generated, the final optimization result obtained by the real-time allocation example does not represent the optimal situation of this study, and is only presented as an example. The real-time allocation results are shown in Table 9.
[0172] Table 9 Real-time allocation results
[0173]
[0174] Comparing the pre-allocation result table 7 with the real-time allocation result table 9, only three flights have changed, including:
[0175] The parking position of flight 1 changes from 213 to 211;
[0176] The parking position of flight 3 is changed from 208 to 213;
[0177] The parking position of flight 7 is changed from 213 to 212.
[0178] Based on the fact that 4 flights experienced minor delays, only 3 flights needed to adjust their parking positions. It can be seen that in the case of minor delays, the real-time allocation using the parking position allocation model constructed in this study only resulted in minor changes in parking positions, proving that the model's real-time allocation in the case of minor delays is effective.
[0179] In addition, since the real-time allocation and pre-allocation use the same model, the optimization objectives are also the same. From the optimization objective of reducing parking space occupancy, the pre-allocated flights are distributed on 8 parking spaces, and the flights after real-time allocation are also distributed on 8 parking spaces. Therefore, it can be seen that the model proposed in this study can reduce the occupancy of parking spaces and increase the efficiency of parking space utilization regardless of whether it is used in pre-allocation or real-time allocation.
[0180] From the perspective of the optimization goal of the shortest taxiing distance, the taxiing distance data (Table 10) of real-time allocation are analyzed. It can be seen that the taxiing distance calculated by pre-allocation is 108268 meters, and the taxiing distance calculated by real-time allocation is 107737 meters. The difference between the results of real-time allocation and pre-allocation is small, which still proves that this model can reduce the taxiing distance by allocating parking positions, which can reduce the taxiing time of aircraft at the airport and reduce the possibility of runway congestion; it can also save a certain amount of taxiing fuel costs for airlines. The parking position allocation model proposed in this invention has certain practical significance.
[0181] Table 10 Real-time allocation optimization results
[0182]
[0183] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for real-time parking space allocation at an airport, characterized in that: include The shortest aircraft taxiing distance and the highest parking space occupancy rate are taken as the optimization objectives, and the safety of the ground operation is taken as the constraint condition to build a real-time parking space allocation model; the function corresponding to the optimization objective is: f=f1+f2 (1) f1=min[∑ i∈N ∑ k∈M x ik (Yes k +DD K )] (2) in, N represents the flight set, i, j∈N; i=1,2,3…20; j=2,3,4…20; M represents the parking space set, k∈M; k=1,2,3…28; DA represents the flight sliding distance set, DA k ∈DA;k∈M;k=1,2,3…28; DD represents the flight taxiing distance set, DD k ∈DD; k∈M; k=1,2,3…28; Among them, f1 represents the optimization goal of the shortest taxiing distance, and the objective function is to find the minimum sum of the taxiing distance and the taxiing distance of each flight, that is, the shortest total taxiing distance; f2 means maximizing the efficiency of parking space utilization and reducing parking space occupancy; The constraints are specifically: in, E ik represents the off-block time for flight i to enter parking stand k; L ik represents the off-block time of flight i leaving parking stand k; E jk represents the block time of flight j that uses the same parking stand k after flight i; Among them, formula (4) constrains that a flight can only be arranged at one parking position; formula (5) indicates that a parking position can only accommodate one flight at most in the same time period; formula (6) indicates that there needs to be a 35-minute safety time interval between two adjacent flights; formula (7) indicates that each flight must stay at the parking position for at least one hour, which is the constraint of the minimum stopover time; formulas (8) and (9) indicate that the airport's operating hours are from 5:00 to 24:00 every day, and flights are only allowed to take off and land within this time range; The parking stand allocation model is optimized using genetic algorithm to obtain the parking stand pre-allocation result.
2. The method for real-time airport parking space allocation according to claim 1, characterized in that: The optimization calculation of the parking space allocation model using a genetic algorithm to obtain the parking space pre-allocation result specifically includes the following: Step 1: Gene encoding using the taxiing distance of each parking position; Step 2: Generate the initial population randomly; Step 3: Set the individual fitness function in the population; Step 4: Obtain the optimal solution through selection, crossover, and mutation calculation, which is the final result of parking space pre-allocation.
3. A real-time parking space allocation device for an airport, characterized in that: include The model building device is used to construct a real-time parking space allocation model with the shortest aircraft taxiing distance and the highest parking space occupancy rate as the optimization objectives and the safety of the ground operation as the constraint condition; the function corresponding to the optimization objective is f=f1+f2 (1) f1=min[∑ i∈N ∑ k∈M x ik (Yes k +DD K )] (2) in, N represents the flight set, i, j∈N; i=1,2,3…20; j=2,3,4…20; M represents the parking space set, k∈M; k=1,2,3…28; DA represents the flight sliding distance set, DA k ∈DA;k∈M;k=1,2,3…28; DD represents the flight taxiing distance set, DD k ∈DD; k∈M; k=1,2,3…28; Among them, f1 represents the optimization goal of the shortest taxiing distance, and the objective function is to find the minimum sum of the taxiing distance and the taxiing distance of each flight, that is, the shortest total taxiing distance; f2 means maximizing the efficiency of parking space utilization and reducing parking space occupancy; The constraints are specifically: in, E ik represents the off-block time for flight i to enter parking stand k; L ik represents the off-block time of flight i leaving parking stand k; E jk represents the block time of flight j that uses the same parking stand k after flight i; Among them, formula (4) constrains that a flight can only be arranged at one parking position; formula (5) indicates that a parking position can only accommodate one flight at most in the same time period; formula (6) indicates that there needs to be a 35-minute safety time interval between two adjacent flights; formula (7) indicates that each flight must stay at the parking position for at least one hour, which is the constraint of the minimum stopover time; formulas (8) and (9) indicate that the airport's operating hours are from 5:00 to 24:00 every day, and flights are only allowed to take off and land within this time range; The model solving device is used to optimize the parking space allocation model by using a genetic algorithm to obtain a parking space pre-allocation result.
4. An electronic device, characterized in that: include at least one processor, and at least one memory in communication with the processor, wherein: The memory stores program instructions executable by the processor, and the processor can execute the method according to any one of claims 1-2 by calling the program instructions.
5. A non-volatile computer-readable storage medium, when the computer-executable instructions are executed by one or more processors, causes the processors to perform the method according to any one of claims 1 to 2.
