摘要
为解决当前航路时空资源分配方法适用场景简单、算法结果不优的问题,研究资源分配优化模型和算法。针对流量受限区混联的复杂拓扑结构,构建了整数规划模型,航班控制采用地面延误、改航和空中延误三种策略组合,优化目标考虑效率和公平性两目标,约束条件设置航班航路唯一指派、航班时隙唯一指派、时隙容量、时隙可执行性和空中最大延误等约束。设计了融合学习机制的改进遗传算法,后代种群通过学习前代个体的遗传经验而快速向帕累托前沿收敛。实验结果表明:相较于传统遗传算法,改进算法能有效提高收敛效率,收敛结果更贴合真实帕累托前沿;相较于现行主流算法,改进算法的满意解能同时提高效率目标15.3%和公平性目标34.3%。
To solve the problems that the application scenario is simple and the result of the algorithm is not satisfactory in air route space-time resource allocation,a model and an algorithm of resource allocation optimization are studied.For the hybrid topological structure of flow constrained areas,an integer programming model is formulated.Flights are controlled by ground delay,rerouting and airborne delay.The objectives consider the efficiency and the equity.The constraints are including the sole assignment of flight to route and flight to slot,slot capacity,slot operability and the maximum airborne delay.Furthermore,an improved genetic algorithm integrated with learning mechanism is designed.The core idea of the improved algorithm is that the offspring populations converge to the Pareto frontier more quickly by learning the mutation experience from previous generations.According to the results,the improved algorithm raises the convergence speed effectively and the output of the improved algorithm is more consistent with the real Pareto front,compared with the traditional genetic algorithm.The satisfactory solution of the improved algorithm simultaneously optimizes the efficiency by 15.3%and the equity by 34.3%,compared with the currently-used algorithm.
作者
郭野晨风
蔡疆
田靖
沈强
丁一波
GUO Ye-chen-feng;CAI Jiang;TIAN Jing;SHEN Qiang;DING Yi-bo(The 28th Research Institute of China Electronics Technology Group Corporation,Nanjing 210000,China;CETC LES Information System Co.,Ltd.,Nanjing 210000,China;Jiangsu Sub-bureau of East China Regional Air Traffic Management Bureau,CAAC,Nanjing 211000,China)
出处
《航空计算技术》
2025年第1期65-70,共6页
Aeronautical Computing Technique
基金
国家重点研发计划项目资助(2022YFB2602404)。
关键词
空中交通
改进遗传算法
时空资源分配
多策略组合
air traffic
improved genetic algorithm
space-time resource allocation
combination of multi traffic management initiatives
