Objective With the steady improvement of China's economy, the civil aviation transportation industry continues to recover, and the increasing demand for travel results in a rapid increase in air traffic flow. However, given the limitation of airspace resources, the increase in the number of aircraft inevitably raises the risk of flight conflicts. Therefore, research on flight conflict resolution becomes crucial, and an efficient and reliable conflict resolution method remains essential to address this issue. Two main shortcomings exist in current research on conflict resolution methods: 1) The commonly used conflict resolution methods exhibit low solution accuracy and slow convergence speed, which limit their applicability across different conflict scenarios. 2) The resolution strategies are overly singular, with most studies focusing solely on heading adjustment without sufficient consideration of composite adjustment strategies that integrate heading and speed. Differential Evolution (DE), a population-based global stochastic search algorithm, possesses a simple structure, strong global optimization performance, robust adaptability, and good scalability, making it a feasible solution for global optimization problems such as flight conflict resolution. However, DE exhibits slow search speed, low convergence accuracy, and a tendency to fall into local optima in multi-aircraft conflict resolution problems, which restricts its practicality. Rat Swarm Optimizer (RSO), a novel heuristic algorithm that simulates rat behaviors of chasing and fighting prey, features a simple structure, few regulation parameters, and fast convergence speed and can effectively compensate for DE's limitations. Based on this, the present study combines the advantages of DE and RSO to propose a hybrid algorithm (HDERSO) and applies it to the problem of flight conflict resolution. Methods HDERSO incorporates the search mechanism of RSO into DE and utilizes a selection probability to adaptively control the generation strategies for offspring individuals at different evolutionary stages. This approach provides directional guidance for the evolution process and avoids blind or ineffective searching. In the early evolutionary stage, the algorithm emphasizes the global exploration capability of DE to search the solution space thoroughly, discover promising regions, and reduce the risk of premature convergence. In the middle and later stages, HDERSO uses the local exploitation ability of RSO to accelerate optimization and enhance convergence accuracy. In addition, control parameters such as the scaling factor and crossover probability influence the quality of DE's offspring individuals. Typically, these parameters require appropriate adjustments during population evolution. Accordingly, HDERSO applies the control parameters to each individual and adaptively generates them by tracking the states of the population and individuals in real-time, improving optimization performance and algorithm applicability. Results and Discussions The performance of HDERSO is evaluated using the IEEE CEC2017 test set comprising 29 functions with diverse features (2 unimodal functions, seven multimodal functions, 10 hybrid functions, and 10 composition functions). The mean and standard deviation of the optimal solutions obtained by HDERSO and the comparative algorithms are statistically analyzed using Wilcoxon and Friedman tests at the 5% significance level. The experimental results demonstrate that the proposed HDERSO achieves higher solution quality and faster convergence speed. Aircraft involved in flight conflict require feasible resolution trajectories based on conflict resolution methods combined with specific strategies, which lead to certain positional offsets from the planned trajectories. Larger positional offsets result in greater flight delays and increased fuel consumption. A flight conflict resolution model based on minimizing the trajectory offset is constructed, and resolution strategies involving heading adjustment, speed adjustment, and composite adjustment are designed. In addition, three typical flight scenarios with varying conflict characteristics (same-direction intersection, vertical intersection, and opposing head-to-head) are developed, and the proposed HDERSO is used in conflict resolution experiments. The results show that HDERSO resolves conflicts using all three strategies in the same-direction intersection scenario. However, in opposing head-to-head scenarios, HDERSO fails to resolve conflicts using speed adjustment alone but successfully resolves them using heading adjustment and composite adjustment. HDERSO effectively resolves conflicts among multiple aircraft by applying different resolution strategies while maintaining the minimum flight safety interval. The results show that the proposed algorithm produces significantly smaller trajectory offsets and target fitness values across various conflict scenarios compared to other algorithms. The evolution curves further confirm that HDERSO is more competitive in convergence speed and solution accuracy, achieving higher accuracy within 250 generations. In addition, the prioritization of the three conflict resolution strategies (heading adjustment, speed adjustment, and composite adjustment) is analyzed based on scenario applicability, solution quality, and operational complexity. The results indicate that in terms of scenario applicability, composite adjustment and heading adjustment outperform speed adjustment; in terms of solution quality, speed adjustment is superior, followed by composite adjustment and heading adjustment; in terms of operational complexity, the strategies are prioritized as heading adjustment, speed adjustment, and composite adjustment. Conclusions Accordingly, the proposed HDERSO is an effective and feasible conflict resolution method that plans optimal resolution trajectories for different flight scenarios, reducing fuel consumption and minimizing flight costs. The outcomes of this study provide a valuable reference for air traffic control departments in implementing flight conflict resolution measures.
WuHaolin.Research on flight conflict resolution strategy based on cooperative multi-agent reinforcement learning[D].Chengdu:Sichuan University,2021.
[2]
吴昊霖.基于协作多智能体强化学习的飞行冲突解脱策略研究[D].成都:四川大学,2021.
[3]
GuanXiangmin, RenliLyu.Conflict resolution method for multiple aircraft based on hybrid artificial potential field and ant colony algorithm[J].Journal of Wuhan University of Technology(Transportation Science & Engineering), 2020,44(1):28‒33.
WuJun, ZhangJingjuan.Conflict resolution of multiple airplanes in free flight based on the genetic algorithm[J].CA- AI Transactions on Intelligent Systems,2013,8(1):16‒20.
FengXingjie, ZhaoRui.Conflict resolution of airplanes ba-sed on multi-objective genetic algorithm[J].Computer Engineering and Design,2014,35(7):2577‒2581.
ZhenRan, WangPan, WuXiaojing,et al.A method of unmanned aerial vehicles conflict resolution based on quantum genetic algorithm[J].Science Technology and Engineering,2020,20(17):6963‒6969.
CuiLiwei, ShiWeiren, LiuXiangming,et al.Air conflict re-solution based on genetic algorithm and particle swarm op-timization[J].Computer Engineering and Applications, 2013,49(7):263‒266.
ZhenRan, SiChao, WuXueli,et al.Study of the aircraft conflict resolution method based on modified particle swarm optimization[J].Journal of Hebei University of Science and Technology,2016,37(5):491‒496.
HongY, ChoiB, OhG,et al.Nonlinear conflict resolution and flow management using particle swarm optimization[J].IEEE Transactions on Intelligent Transportation Syste-ms,2017,18(12):3378‒3387. doi:10.1109/tits.2017.2684824
[16]
XieHua, SuFangzheng, YinJianan,et al.Research on classified detection and differential resolution method for U-AV flight conflicts in low altitude airspace[J].Journal of Safety and Environment,2023,23(9):3131‒3142.
OmerJ.A space-discretized mixed-integer linear model for air-conflict resolution with speed and heading maneuvers[J].Computers and Operations Research,2015,58:75‒86. doi:10.1016/j.cor.2014.12.012
[19]
WangZ, LiH, WangJ,et al.Deep reinforcement learning based conflict detection and resolution in air traffic control[J].IET Intelligent Transport Systems,2019,13(6):1041‒1047. doi:10.1049/iet-its.2018.5357
[20]
GhoshS, LagunaS, LimS H,et al.A deep ensemble me-thod for multi-agent reinforcement learning:A case study on air traffic control[C]//Proceedings of the International Conference on Automated Planning and Scheduling.Gu-angzhou:AAAI,2021,31:468‒476. doi:10.1609/icaps.v31i1.15993
[21]
SuiD, XuW, ZhangK.Study on the resolution of multi-aircraft flight conflicts based on IDQN[J].Chinese Journal of Aeronautics,2022,35(2):195‒213. doi:10.1016/j.cja.2021.03.015
[22]
PriceK V, StornR M, LampinenJ A.Differential evolution:A practical approach to global optimization[M].Berlin:Springer,2005. doi:10.1007/978-3-540-39930-8_6
[23]
AgrawalS, TiwariA.Solving multimodal optimization pro-blems using adaptive differential evolution with archive[J].Information Sciences,2022,612:1024‒1044. doi:10.1016/j.ins.2022.09.023
[24]
Abdel‒NabiH, AliM Z, AwajanA,et al.An iterative cyclic tri-strategy hybrid stochastic fractal with adaptive differential algorithm for global numerical optimization[J].Information Sciences,2023,628:92‒133. doi:10.1016/j.ins.2023.01.065
[25]
AliI M, EssamD, KasmarikK.Novel binary differential evolution algorithm for knapsack problems[J].Information Sciences,2021,542:177‒194. doi:10.1016/j.ins.2020.07.013
[26]
SallamK M, AbohanyA A, Rizk‒AllahR M.An enhanced multi-operator differential evolution algorithm for tackling knapsack optimization problem[J].Neural Computing and Applications,2023,35:13359‒13386. doi:10.1007/s00521-023-08358-7
PlichoskiG F, ChidambaramC, ParpinelliR S.A face recognition framework based on a pool of techniques and differential evolution[J].Information Sciences,2021,543:219‒ 241. doi:10.1016/j.ins.2020.06.054
[30]
WuB, WangL, LvS,et al.Forecasting oil consumption with attention-based IndRNN optimized by adaptive differential evolution[J].Applied Intelligence,2023,53:5473‒5496. doi:10.1007/s10489-022-03720-z
[31]
GaoZ, ZhangM, ZhangL.Ship-unloading scheduling optimization with differential evolution[J].Information Sciences,2022,591:88‒102. doi:10.1016/j.ins.2021.12.110
WangY, LiuZ, WangG.Improved differential evolution using two-stage mutation strategy for multimodal multi-obj-ective optimization[J].Swarm and Evolutionary Computation,2023,78:101232. doi:10.1016/j.swevo.2023.101232
[35]
DhimanG, GargM, NagarA,et al.A novel algorithm for global optimization:Rat swarm optimizer[J].Journal of Am-bient Intelligence and Humanized Computing,2021,12:8457‒8482. doi:10.1007/s12652-020-02580-0
[36]
XuZiyue, LiangXiaodan.Opposition-based learning DE-RSO hybrid optimizer[J].Computer Science and Application,2021,11(12):2890‒2899.
AwadN H, AliM Z.Liang J J,et al.Problem definitions and evaluation criteria for the CEC 2017 special session and competition on single objective bound constrained real-parameter numerical optimization[R].Singapore:Nanyang Technological University,2016.