航班着陆管制中结合扰动算子与参考集的改进VNS优化方法

An Improved VNS Algorithm Combined with Disturbing Operator and Reference Setfor Aircraft Landing Scheduling Optimization

飞机着陆问题(ALP)的合理调度对维持飞行安全和最小化航班延误非常重要。提出了一种改进型变邻域搜索(VNS)算法,来求解ALP优化模型。首先,在邻域搜索过程中,根据分散搜索算法中的参考集思想,用包含具备高质量且分散性的参考集代替VNS中的初始邻域解集,以此减少无效搜索。其次,对局部搜索失败次数进行计数,通过融入一个扰动算子来帮助算法跳出局部最优陷阱。最后,构建一个收敛精度高且速度快的改进型VNS算法。实验结果表明,该方法能够合理调度航班着陆,最小化延误成本。

Reasonable solution of aircraft landing problem(ALP)is very important for maintaining flight safety and minimizing flight delays.An improved variable neighborhood search(VNS)algorithm is proposed to solve the ALP optimization model.Firstly,in the process of neighborhood search,according to the idea of reference set in scatter search algorithm,the initial neighborhood solution set in VNS is replaced by a reference set with high quality and decentralization,so as to reduce invalid search.Then,the number of local search failures is counted and a disturbing operator is incorporated to help it escape from the local optimal trap.Finally,an improved VNS algorithm with high convergence accuracy and fast speed is constructed.The experimental results in some ALP instances show that the method can reasonably schedule the flight landing and minimize the delay cost.