An airport parking space adaptive multi-objective intelligent allocation system and method

Abstract

The present application relates to the technical field of air traffic management and intelligent optimization, and particularly relates to an airport parking space adaptive multi-objective intelligent allocation system and method, which comprises a heuristic initialization module, an adaptive parameter control module, a constraint-aware genetic operation module and a comprehensive decision output module. The final allocation scheme is obtained by inputting basic data into the corresponding modules. Specifically, the initial population is generated by using heuristic rules to meet the constraints of aircraft type matching and time conflict-free. Then, the parameter adaptive adjustment is performed by dynamically adjusting the crossover and mutation probabilities according to the population diversity. Through constraint crossover, feasible neighborhood mutation and greedy repair, the multi-objective intelligent decision is made, and the executable scheme is directly output by normalization and weighted scoring. The present application effectively breaks through the bottleneck of traditional algorithms, can significantly improve the utilization rate of airport parking space resources, reduce operating costs and carbon emissions, and improve passenger service experience, thereby providing efficient and feasible technical support for airport intelligent scheduling and green operation.

Description

Technical Field

[0001] This invention relates to the field of interdisciplinary technology of air traffic management and intelligent optimization, specifically to an adaptive multi-objective intelligent allocation system and method for airport parking positions. Background Technology

[0002] Aircraft parking stands are one of the most critical and scarce ground resources at an airport. Proper allocation of parking stands not only enhances the passenger experience but also optimizes aircraft taxiing paths, reducing fuel consumption and carbon emissions. It is a key element in improving airport operational efficiency and achieving the goal of a "green airport." The Gate Assignment Problem (GAP) is essentially a dynamic combinatorial optimization problem with multiple constraints and objectives.

[0003] Currently, research on parking space allocation problems mainly falls into three categories. The first category is based on deterministic rules, such as first-come, first-served, which is simple but inefficient and struggles to cope with dynamic changes during flight operations. The second category is single-objective optimization methods, typically aiming to minimize passenger walking distance or maximize jet bridge utilization. However, in actual operations, airlines and airport management often face multiple conflicting objectives (such as the trade-off between high jet bridge utilization and low taxiing fuel consumption), making single-objective optimization insufficient to fully reflect operational needs. The third category is based on multi-objective evolutionary algorithms, such as the NSGA-II algorithm, which can generate a set of Pareto optimal solutions.

[0004] However, traditional NSGA-II still faces significant bottlenecks when dealing with the stop position allocation problem with complex constraints: 1) Limitations of fixed evolutionary parameters: Traditional algorithms typically use fixed crossover and mutation probabilities, which cannot be adaptively adjusted according to the population's evolutionary state, easily leading to premature convergence or slow convergence speed. 2) Insufficient constraint handling capability: Traditional genetic operators do not fully consider the specific constraints of the stop position allocation problem, and the generated offspring solutions often violate the constraints. Existing repair mechanisms are inefficient, resulting in a low proportion of feasible solutions and limited search efficiency. 3) Poor initial population quality: Traditional genetic algorithms use random initialization to generate the initial population, and a large number of individuals violate the constraints, causing early iterations to be wasted in the invalid search space, affecting convergence efficiency. 4) Disconnect between objective trade-offs and decision-making: Multi-objective optimization algorithms output a Pareto solution set, rather than a single execution scheme. In engineering practice, there is a lack of effective mechanisms to connect the selection of an "optimal" scheme from the solution set that satisfies operational preferences and is easy to implement. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive multi-objective intelligent allocation system and method for airport parking positions, aiming to solve the technical problems of traditional parking position allocation methods, such as fixed parameters leading to premature convergence, insufficient constraint handling capabilities, low quality of initial solutions, and difficulty in applying multi-objective results.

[0006] To achieve the above objectives, the present invention provides an adaptive multi-objective intelligent allocation system for airport parking positions, including a heuristic initialization module, an adaptive parameter control module, a constraint-aware genetic operation module, and a comprehensive decision output module. Data is sequentially processed through the heuristic initialization module, the adaptive parameter control module, the constraint-aware genetic operation module, and the comprehensive decision output module to output a parking position allocation scheme.

[0007] The heuristic initialization module constructs a feasible set of machine slots and generates an initial population by combining greedy selection with random perturbation.

[0008] The adaptive parameter control module uses a sliding weighted average method to construct a dynamic threshold and adaptively adjusts the crossover probability and mutation probability according to the population diversity status.

[0009] The constraint-aware genetic operation module is responsible for improving the adaptability of genetic operations to complex constraints in stop position allocation;

[0010] The comprehensive decision output module transforms the non-dominated solution set into an directly executable allocation scheme through weighted comprehensive evaluation calculation.

[0011] Furthermore, this invention also proposes an adaptive multi-objective intelligent allocation method for airport parking stands, employing the aforementioned adaptive multi-objective intelligent allocation system for airport parking stands, comprising the following steps:

[0012] Step 1: Use the heuristic initialization module to preprocess the basic data and generate the initial population;

[0013] Step 2: The adaptive parameter control module dynamically adjusts the crossover and mutation probabilities by monitoring the distribution characteristics of the population in real time.

[0014] Step 3: The constraint-aware genetic operation module performs two-point crossover, adaptive mutation, and greedy repair operations to ensure that the offspring population always maintains high feasibility.

[0015] Step 4: Obtain the Pareto optimal solution set based on the multi-objective optimization algorithm. The comprehensive decision output module transforms the Pareto optimal solution set into a single executable solution through range normalization and weighted scoring, and outputs the allocation result.

[0016] Optionally, the basic data includes flight information and parking position information, and the execution process of step 1 includes the following steps:

[0017] Step 1.1: Flight sorting, which sorts all flights to be assigned in ascending order of time based on their scheduled arrival time or the estimated arrival time entered by the system.

[0018] Step 1.2: Assign parking positions to each flight in the sorted order to form a feasible allocation set (Candidate). i ;

[0019] Step 1.3: From the feasible allocation set Candidate i Randomly select a parking position as the allocation result for the flight and update the occupancy status table of the corresponding parking position; repeat the process until all flights are allocated, generating a complete allocation scheme that satisfies all constraints; repeat the process from step 1.1 to step 1.3 N times, where N is the population size, to generate an initial population consisting entirely of feasible solutions.

[0020] Optionally, the stopping position in step 1.2 satisfies the following hard constraints:

[0021] Aircraft type-gate compatibility constraint, which means that the type of aircraft used for a flight must match the type of parking stand;

[0022] Time exclusivity constraints ensure that a parking stand can serve a maximum of one flight at any given time, and that a minimum safe time interval must be maintained between adjacent flights.

[0023] Optionally, the execution process of step 2 includes the following steps:

[0024] Step 2.1: Use the crowding distribution on the Pareto front as a diversity indicator to quantify the dispersion of the current population in the target space;

[0025] Step 2.2: Construct a dynamic threshold using the moving weighted average method;

[0026] Step 2.3: Adaptively adjust the crossover probability and mutation probability based on the dynamic threshold.

[0027] Optionally, the execution process of step 3 includes the following steps:

[0028] Step 3.1: Randomly select two crossover points in the flight assignment sequence of the parent individuals and exchange the gene segments of the two parent individuals within the crossover interval;

[0029] Step 3.2: Introduce feasible neighborhood constraints within feasible machine locations that are compatible with the machine type and do not conflict with the time.

[0030] Step 3.3: Execute the greedy repair procedure immediately after each crossover or mutation operation.

[0031] Optionally, the multi-objective optimization algorithm in step 4 includes two objective functions: maximizing the flight docking rate and minimizing the flight arrival taxiing fuel consumption cost.

[0032] Optionally, the execution process of step 4 includes the following steps:

[0033] Step 4.1: Perform range normalization on the objective value of each solution;

[0034] Step 4.2: Set weighting coefficients and perform a weighted comprehensive evaluation.

[0035] This invention provides an adaptive multi-objective intelligent allocation system and method for airport parking positions. The system includes a heuristic initialization module, an adaptive parameter control module, a constraint-aware genetic operation module, and a comprehensive decision output module. The final allocation scheme is obtained by inputting basic data into the corresponding modules. Specifically, an initial population is first generated using heuristic rules to satisfy constraints such as aircraft type matching and time conflict-free operation. Then, the crossover and mutation probabilities are dynamically adjusted based on population diversity to adaptively regulate parameters. Through constraint crossover, feasible neighborhood mutation, and greedy repair, and after multi-objective intelligent decision-making, an executable scheme is directly output after normalization and weighted scoring. This invention effectively overcomes the bottlenecks of traditional algorithms, significantly improving the utilization rate of airport parking position resources, reducing operating costs and carbon emissions, and improving passenger service experience, providing efficient and feasible technical support for intelligent airport scheduling and green operation. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the overall architecture of an adaptive multi-objective intelligent allocation system for airport parking positions according to the present invention.

[0038] Figure 2 This is a schematic diagram comparing the bridge passability results of three methods in a specific embodiment of the present invention.

[0039] Figure 3 This is a schematic diagram comparing the fuel consumption results of three methods in a specific embodiment of the present invention.

[0040] Figure 4 This is a schematic diagram comparing the comprehensive scores of the three methods in a specific embodiment of the present invention.

[0041] Figure 5 This is a schematic diagram showing the parking position allocation result using the method of this invention in a specific embodiment of the invention.

[0042] Figure 6 This is a schematic diagram of the parking position allocation result using the NSGA-II method in a specific embodiment of the present invention.

[0043] Figure 7 This is a schematic diagram of the actual parking space allocation results in a specific embodiment of the present invention. Detailed Implementation

[0044] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0045] This invention provides an adaptive multi-objective intelligent allocation system for airport parking positions, including a heuristic initialization module, an adaptive parameter control module, a constraint-aware genetic operation module, and a comprehensive decision output module. Data is sequentially processed through the heuristic initialization module, the adaptive parameter control module, the constraint-aware genetic operation module, and the comprehensive decision output module to output a parking position allocation scheme.

[0046] The heuristic initialization module constructs a feasible set of machine slots and generates an initial population by combining greedy selection with random perturbation.

[0047] The adaptive parameter control module uses a sliding weighted average method to construct a dynamic threshold and adaptively adjusts the crossover probability and mutation probability according to the population diversity status.

[0048] The constraint-aware genetic operation module is responsible for improving the adaptability of genetic operations to complex constraints in stop position allocation;

[0049] The comprehensive decision output module transforms the non-dominated solution set into an directly executable allocation scheme through weighted comprehensive evaluation calculation.

[0050] Furthermore, this invention also proposes an adaptive multi-objective intelligent allocation method for airport parking stands, employing the aforementioned adaptive multi-objective intelligent allocation system for airport parking stands, comprising the following steps:

[0051] Step 1: Use the heuristic initialization module to preprocess the basic data and generate the initial population;

[0052] Step 2: The adaptive parameter control module dynamically adjusts the crossover and mutation probabilities by monitoring the distribution characteristics of the population in real time.

[0053] Step 3: The constraint-aware genetic operation module performs two-point crossover, adaptive mutation, and greedy repair operations to ensure that the offspring population always maintains high feasibility.

[0054] Step 4: Obtain the Pareto optimal solution set based on the multi-objective optimization algorithm. The comprehensive decision output module transforms the Pareto optimal solution set into a single executable solution through range normalization and weighted scoring, and outputs the allocation result.

[0055] For details, please refer to Figure 1 , Figure 1 The following is a schematic diagram of the overall architecture and process of the system of the present invention, and further explanation is given in conjunction with the execution process of each module and method:

[0056] 1. Heuristic initialization module

[0057] To address the problem of traditional genetic algorithms generating numerous infeasible solutions through random initialization, this module employs heuristic rules to generate the initial population, ensuring that every individual in the population is a feasible solution from the very beginning of evolution. The core steps of Step 1 are as follows:

[0058] Step 1: Flight Sort. Sort all flights to be assigned in ascending order of their scheduled arrival time (or the estimated arrival time entered by the system).

[0059] Step 2: Construct a feasible allocation set. Assign parking positions to each flight in sorted order. For the current flight... Traverse all parking positions Select stop positions that satisfy the following hard constraints to form a feasible allocation set (Candidate). i :

[0060] ① Aircraft type-gate compatibility constraints: The aircraft type of the flight must match the type of parking stand. Large parking stands can accommodate all aircraft types, medium parking stands can accommodate medium and small aircraft types, and small parking stands can only accommodate small aircraft types.

[0061] ② Time exclusivity constraint: For each candidate stop position Check flights Expected occupancy period Does this overlap with the occupancy period of any flights already assigned to that parking position? It must be ensured that at any given time, a parking position can serve a maximum of one flight, and a minimum safe interval must be maintained between adjacent flights. .

[0062] Step 3: Assign individual. Randomly select a parking space from the feasible allocation set Candidatei as the allocation result for this flight, and update the parking space. The occupancy status table is used. This process is repeated until all flights are allocated, generating a complete allocation scheme that satisfies all constraints. The above process is then repeated. Second-rate( (where the population size is ), generate an initial population consisting entirely of feasible solutions.

[0063] Through the above steps, all individuals in the generated initial population are feasible solutions, providing a high-quality starting point for subsequent evolutionary searches.

[0064] 2. Adaptive Parameter Control Module

[0065] To balance the algorithm's global exploration and local exploitation capabilities, this invention employs a diversity-aware adaptive parameter control mechanism in step 2. This mechanism dynamically adjusts the crossover and mutation probabilities by monitoring the population's distribution characteristics in real time, as detailed below:

[0066] Step 1: Quantify diversity status. Use the crowding distribution on the Pareto front as a diversity indicator to quantify the dispersion of the current population in the target space. The higher the calculated crowding, the better the population diversity.

[0067] Step 2: Dynamic threshold construction. A dynamic threshold is constructed using a moving weighted average method. When the measured population diversity is lower than this threshold, it indicates that the population may be trapped in a local optimum or experiencing premature convergence.

[0068] Step 3: Adaptive parameter adjustment. (1) When diversity is below the threshold, increase the crossover and mutation probabilities to enhance global exploration capabilities and help the population escape local optima. (2) When diversity is above the threshold, decrease the crossover and mutation probabilities to enhance local exploration capabilities and improve convergence accuracy. The adaptive adjustment range is preset to: Cross_range[0.5, 0.8], Mut_range[0.05, 0.3]. This mechanism enables the algorithm to automatically adjust the search intensity according to the population state, avoiding the performance bottleneck caused by fixed parameters.

[0069] 3. Constraint-aware genetic operation module

[0070] To improve the adaptability of genetic operators to the stopping position allocation problem, a set of constraint-aware genetic operation mechanisms is used in step 3, including a two-point crossover operator, an adaptive mutation operator, and a greedy repair strategy. The specific process is as follows:

[0071] (1) Two-point crossover operator: Two crossover points are randomly selected in the flight allocation sequence of the parent individuals, and the gene segments (i.e., the seat allocation sequence) of the two parent individuals in the crossover interval are exchanged. This operator can introduce new gene combinations while retaining some high-quality gene combinations, thereby enhancing the search capability.

[0072] (2) Adaptive Mutation Operator: For the mutation operation, this invention introduces a feasibility neighborhood constraint. For the selected flight, its mutation operation is limited to the set of candidate seats that meet the following conditions:

[0073] ① Space compatibility: The gate must be compatible with the aircraft type of the flight.

[0074] ② No time conflict: During the time window occupied by this flight, the gate is not occupied by other flights and maintains the minimum safe time interval with the preceding and following flights.

[0075] In this invention, the mutation operation is performed only within the aforementioned feasible neighborhood, ensuring that the mutated solution still satisfies the basic constraints.

[0076] (3) Greedy Repair Strategy: After each crossover or mutation operation, a greedy repair procedure is executed immediately. This procedure checks the allocation of each flight in chronological order to ensure that it meets the aircraft type matching and time window constraints. For any allocation that violates the constraints, the procedure selects the optimal aircraft stand from the set of feasible stands for that flight and reassigns it, updating the stand occupancy status in real time until all flight allocations meet the constraints. This repair strategy ensures that the offspring population always maintains high feasibility.

[0077] 4. Comprehensive Decision Output Module

[0078] To meet the combined needs of passenger satisfaction, airport operating costs, and emissions reduction, this invention proposes the following two objective functions:

[0079] (1) Maximize the flight gate utilization rate

[0080] Maximizing gate utilization reflects the core value of improving aviation service quality and passenger satisfaction. Allocating flights to gates closer to boarding bridges not only reduces passenger commuting time and increases passenger satisfaction, but also reduces the demand for shuttle buses, thereby lowering airport operating costs. In practice, this objective translates to maximizing the number of flights allocated to boarding bridge gates. Therefore, the objective function for the first objective is shown in the equation.

[0081]

[0082] (2) Minimize the fuel consumption cost of flights taxiing to the port.

[0083] Controlling taxiing fuel consumption during the ground phase reflects the dual demands of economic efficiency and environmental sustainability in aviation operations. From the perspective of airline cost control, optimizing parking stand allocation to reduce ground taxiing distance can significantly reduce fuel consumption, thereby generating direct economic benefits. From an environmental engineering perspective, fuel combustion during aircraft ground taxiing is one of the main sources of air pollution in airport areas. Combining these two perspectives, it can be seen that optimizing arrival taxiing fuel consumption not only reduces operating costs but also aligns with the sustainable development requirements of the CORSIA (Carbon Offset and Reduction in International Aviation) program, embodying the core principles of green airport construction. Therefore, the following second objective function is proposed:

[0084]

[0085] The multi-objective optimization algorithm ultimately outputs a set of Pareto optimal solutions. To facilitate direct decision-making by airport operators, step 4 of this invention employs a comprehensive decision output module to map the multi-objective solution set into a single executable solution. The core steps of the execution process include:

[0086] Step 1: Objective value normalization. Since the two objective functions have different dimensions and orders of magnitude, direct weighted summation is meaningless. Therefore, the objective value of each solution is normalized to its range:

[0087]

[0088]

[0089] Among them, utilization rate is the goal of maximization, and fuel consumption is the goal of minimization. After normalization, both are in the form of maximization, and their value range is [0,1].

[0090] Step 2: Weighted Overall Evaluation. Based on the airport's actual operational preferences (such as prioritizing passenger experience or minimizing operating costs), weighting coefficients are set, and the overall score for each solution is calculated:

[0091]

[0092] Finally, select the highest-scoring solution and output the final allocation.

[0093] For further details, please refer to Figures 2 to 7 The present invention also proposes specific embodiments, which are illustrated by comparison with existing methods:

[0094] 1. Data Preparation: Forty flight operation data points from Zhengzhou Xinzheng International Airport (CGO) were selected as the allocation sample, including basic information such as flight number, aircraft type, Actual Arrival Time (ATA), and Actual Departure Time (ATD). Parking information included the type of 10 parking stands (classes 1-3), whether they were jet bridge stands, etc., as shown in Tables 1 and 2 below.

[0095] Table 1 Basic Flight Information

[0096]

[0097] Table 2 Basic Information of Parking Positions

[0098]

[0099] 2. Problem Constraints: To ensure airport operational safety and the feasibility of optimizing parking stand allocation, all parking stands allocated to flights must meet strict constraints, as follows:

[0100] (1) Exclusivity constraint: Ensure that each flight must be assigned one and only one parking position.

[0101]

[0102] (2) Aircraft stand space compatibility constraints: Ensure that the number of flights served by each parking stand cannot exceed its space capacity limit, and that flights can only be assigned to parking stands that are compatible with aircraft type. Among them, large parking stands are applicable to all aircraft types, medium parking stands are applicable to small and medium-sized aircraft types, and small parking stands are only applicable to small aircraft types.

[0103]

[0104]

[0105] (3) Safety time interval constraint: Ensure that each parking stand serves at most one flight at any time point t, and for any two adjacent flights on the same parking stand j and A minimum safe time interval must be maintained between flights to ensure that there are no time conflicts between them.

[0106]

[0107]

[0108]

[0109] (4) Uniqueness constraint of connecting flights: Ensure that each flight has at most one direct preceding flight and one following flight at the same gate.

[0110]

[0111]

[0112] (5) Symmetry constraint: In addition to the uniqueness of connecting flights, symmetry verification is added, that is, the adjacent relationships between flights are mutually exclusive in direction.

[0113]

[0114] The specific meanings of the variables used are shown in Table 3.

[0115] Table 3. Variable Definitions for Parking Position Allocation Model

[0116]

[0117] 3. Flight sorting: Sort the 40 flights in ascending order of their scheduled arrival times to form a sequence to be allocated.

[0118] 4. Construct and allocate feasible stands for each flight: For each sorted flight, construct a feasible stand set based on its aircraft type and current stand occupancy status. Feasible stands must meet the following requirements: ① Aircraft type compatibility (medium-sized aircraft can park at Class 2 and 3 stands); ② No time window conflicts (interval with the occupancy window of an already allocated flight is ≥ 15 minutes from the minimum safe interval). Allocation is performed from the feasible set using a greedy strategy (prioritizing jet bridge stands), and the stand occupancy status is updated in real time.

[0119] 5. Initial Population Generation: Repeat the above process, introducing a small amount of random perturbation to generate an initial population of size 60. Verification shows that all initial individuals satisfy the hard constraints, and the feasible solution rate reaches 100%.

[0120] The following uses the method of this invention (abbreviated as: ADA-MOGA) for adaptive diversity multi-objective parking space allocation optimization:

[0121] 1) Adaptive Parameter Control: This mechanism dynamically adjusts the intensity of genetic operations by monitoring population distribution characteristics in real time. It uses the crowding distribution at the Pareto front as a diversity metric, calculates the dispersion of the current population distribution in the target space, and establishes a dynamic threshold baseline using a moving weighted average method. When population diversity is detected to be below the threshold, the crossover and mutation probabilities are adaptively increased to enhance global exploration capabilities and avoid premature convergence; conversely, they are appropriately reduced to strengthen local development and improve convergence accuracy. Specifically, the initial crossover probability is set to 0.6 and the mutation probability to 0.15. In each iteration, the crowding of the population at the Pareto front is calculated. When the crowding is below the moving average threshold, the crossover probability is dynamically increased to 0.8 and the mutation probability to 0.3 to enhance exploration capabilities; conversely, the crossover probability is decreased to 0.5 and the mutation probability to 0.05 to strengthen development capabilities.

[0122] 2) Constrained sensing genetic operations:

[0123] Crossover: A two-point crossover is used, where flight sequences within the crossover interval are exchanged to introduce new gene combinations while retaining some superior patterns.

[0124] Mutation: For the selected flight, randomly select a new gate from its set of feasible gates (satisfying aircraft type compatibility and time non-conflict) to replace it, ensuring that the mutated solution does not violate the constraints.

[0125] Repair: After crossover or mutation, a greedy repair strategy is invoked to check the allocation of each flight in chronological order. For illegal allocations, the optimal gate (priority jet bridge gate) is selected from the set of feasible gates and reassigned.

[0126] 3) Multi-objective optimization: With the goal of maximizing the utilization rate of the boardwalk and minimizing the taxiing fuel consumption, the algorithm was run for 95 iterations with a population size of 60.

[0127] 4) Comprehensive decision-making output

[0128] 1. Pareto solution set generation: Run the ADA-MOGA algorithm to obtain a set of non-dominated solutions, i.e., the Pareto front.

[0129] 2. Range normalization: Normalize the range of the bridge utilization (U) and coasting fuel consumption (F) of all solutions on the frontier so that they are all in the [0,1] interval and are all in the maximum form.

[0130] 3. Weighted scoring and scheme selection: Based on the airport's operational preferences, the weight of the jet bridge utilization rate is set as w1 = 0.65, and the weight of taxiing fuel consumption is set as w2 = 0.35. The comprehensive score of each solution is calculated, and the solution with the highest score is selected as the final execution scheme.

[0131] The final comparative experiment and results verification are as follows:

[0132] Comparison Group 1: Traditional NSGA-II (using fixed parameters, unconstrained perceptual operator)

[0133] Comparison Group 2: Actual Airport Allocation Scheme

[0134] The comparison results show the bridge berth rate for example. Figure 2 As shown, the fuel consumption results are compared in the following figures. Figure 3 The overall score is as follows: Figure 4 As shown.

[0135] Results analysis:

[0136] Bridge utilization rate: The method of this invention (92.5%) > NSGA-II (82.5%) > actual (80%).

[0137] Coasting fuel consumption: The method of this invention (594.1×10³L) < Actual (602.7×10³L) < NSGA-II (717.4×10³L).

[0138] Overall score: The method of this invention (92.85) > NSGA-II (79.58).

[0139] Gantt chart comparison: Figures 5 to 7 The actual allocation results of the three schemes are shown. Figure 5 As shown, the method of the present invention maintains two empty gates (8 and 9) while achieving a more compact occupancy rate, which helps to arrange flights with longer stop times, thereby improving the flexibility of resource utilization and providing flexibility for long turnaround flights.

[0140] The experimental results fully demonstrate the significant advantages of this invention in improving the utilization rate of the boardwalk, reducing taxiing fuel consumption, and achieving an effective balance in operational objectives.

[0141] In summary, the present invention has the following beneficial effects:

[0142] 1. Improve the quality of the initial population and search efficiency.

[0143] The heuristic initialization method ensures that all individuals satisfy machine compatibility and time window constraints when generating the initial population, avoiding a large amount of invalid search caused by traditional random initialization. Experiments show that after adopting heuristic initialization, the algorithm can obtain high-quality feasible solutions in the early iterations, and the convergence speed is significantly improved.

[0144] 2. Balancing global exploration with local development

[0145] The diversity-aware adaptive parameter control mechanism can monitor population crowding in real time and dynamically adjust crossover and mutation probabilities. When population diversity decreases, it actively increases the mutation probability to effectively avoid premature convergence; when population diversity is abundant, it decreases the mutation probability to accelerate convergence.

[0146] 3. Improve the efficiency and quality of feasible solution generation.

[0147] The constraint-aware two-point crossover operator enhances search capabilities while preserving high-quality gene combinations. The feasibility neighborhood mutation operator restricts mutation operations to the set of feasible aircraft positions that satisfy aircraft type compatibility and time conflict-free conditions, ensuring the feasibility of the mutated solution. The greedy repair strategy acts as the final barrier, efficiently repairing the few remaining potentially illegal allocations, ensuring the entire search process operates efficiently within the feasible solution space. Experiments show that the final allocation scheme achieves a 100% conflict-free flight ratio.

[0148] 4. Effective trade-offs and decision support for achieving operational goals

[0149] This invention's algorithm generates a set of Pareto optimal solutions, clearly demonstrating the trade-off between jet bridge utilization and taxiing fuel consumption. Then, through range normalization and weighted scoring, the multi-objective solution set is transformed into a single executable solution. With a weight of w1 = 0.65, the algorithm achieves a comprehensive score of 92.85, significantly higher than the traditional NSGA-II (79.58). In terms of actual operational benefits, in a case study of 40 flights and 10 aircraft stands, the algorithm achieved a jet bridge utilization rate of 92.5%, a 12.5% ​​improvement over the actual operating plan (80%); taxiing fuel consumption was 594.1 × 10³ L, a 1.43% reduction compared to the actual plan (602.7 × 10³ L). This improves the passenger experience while also bringing environmental benefits.

[0150] The above description discloses only one or more preferred embodiments of the present invention, and should not be construed as limiting the scope of the present invention. Those skilled in the art will understand that all or part of the processes of the above embodiments can be implemented, and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. An adaptive multi-objective intelligent allocation system for airport parking positions, characterized in that, It includes a heuristic initialization module, an adaptive parameter control module, a constraint-aware genetic operation module, and a comprehensive decision output module. The data passes through the heuristic initialization module, the adaptive parameter control module, the constraint-aware genetic operation module, and the comprehensive decision output module in sequence before the parking position allocation scheme is output. The heuristic initialization module constructs a feasible set of machine slots and generates an initial population by combining greedy selection with random perturbation. The adaptive parameter control module uses a sliding weighted average method to construct a dynamic threshold and adaptively adjusts the crossover probability and mutation probability according to the population diversity status. The constraint-aware genetic operation module is responsible for improving the adaptability of genetic operations to complex constraints in stop position allocation; The comprehensive decision output module transforms the non-dominated solution set into an directly executable allocation scheme through weighted comprehensive evaluation calculation.

2. An adaptive multi-objective intelligent allocation method for airport parking stands, employing the adaptive multi-objective intelligent allocation system for airport parking stands as described in claim 1, characterized in that, Includes the following steps: Step 1: Use the heuristic initialization module to preprocess the basic data and generate the initial population; Step 2: The adaptive parameter control module dynamically adjusts the crossover and mutation probabilities by monitoring the distribution characteristics of the population in real time. Step 3: The constraint-aware genetic operation module performs two-point crossover, adaptive mutation, and greedy repair operations to ensure that the offspring population always maintains high feasibility. Step 4: Obtain the Pareto optimal solution set based on the multi-objective optimization algorithm. The comprehensive decision output module transforms the Pareto optimal solution set into a single executable solution through range normalization and weighted scoring, and outputs the allocation result.

3. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 2, characterized in that, The basic data includes flight information and parking position information. The execution process of step 1 includes the following steps: Step 1.1: Flight sorting, which sorts all flights to be assigned in ascending order of time based on their scheduled arrival time or the estimated arrival time entered by the system. Step 1.2: Assign parking positions to each flight in the sorted order to form a feasible allocation set (Candidate). i ; Step 1.3: From the feasible allocation set Candidate i Randomly select a parking position as the allocation result for the flight and update the occupancy status table of the corresponding parking position; repeat the process until all flights are allocated, generating a complete allocation scheme that satisfies all constraints; Repeat steps 1.1 to 1.3 N times, where N is the population size, to generate an initial population consisting entirely of feasible solutions.

4. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 3, characterized in that, The stopping position in step 1.2 satisfies the following hard constraints: Aircraft type-gate compatibility constraint, which means that the type of aircraft used for a flight must match the type of parking stand; Time exclusivity constraints ensure that a parking stand can serve a maximum of one flight at any given time, and that a minimum safe time interval must be maintained between adjacent flights.

5. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 4, characterized in that, The execution process of step 2 includes the following steps: Step 2.1: Use the crowding distribution on the Pareto front as a diversity indicator to quantify the dispersion of the current population in the target space; Step 2.2: Construct a dynamic threshold using the moving weighted average method; Step 2.3: Adaptively adjust the crossover probability and mutation probability based on the dynamic threshold.

6. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 5, characterized in that, The execution process of step 3 includes the following steps: Step 3.1: Randomly select two crossover points in the flight assignment sequence of the parent individuals and exchange the gene segments of the two parent individuals within the crossover interval; Step 3.2: Introduce feasible neighborhood constraints within feasible machine locations that are compatible with the machine type and do not conflict with the time. Step 3.3: Execute the greedy repair procedure immediately after each crossover or mutation operation.

7. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 6, characterized in that, The multi-objective optimization algorithm described in step 4 includes two objective functions: maximizing the flight docking rate and minimizing the flight arrival taxiing fuel consumption cost.

8. The airport parking stand adaptive multi-objective intelligent allocation method as described in claim 7, characterized in that, The execution process of step 4 includes the following steps: Step 4.1: Perform range normalization on the objective value of each solution; Step 4.2: Set weighting coefficients and perform a weighted comprehensive evaluation.

Images