基于变邻域搜索改进的冲突解脱粒子群算法  

Improved particle swarm optimization for flight conflict resolution based on variable neighborhood search

作　　者：马兰[1] 李伟岸[1] 尹天懿 

机构地区：[1]中国民航大学空中交通管理学院,天津300300

出　　处：《山东大学学报（理学版）》2017年第1期23-28,42,共7页Journal of Shandong University(Natural Science)

基　　金：国家自然基金委员会和中国民用航空局联合基金资助项目(U1333116)

摘　　要：飞行冲突解脱是航空器安全运行的关键,粒子群优化(particle swarm optimization,PSO)算法和变邻域搜索(variable neighborhood search,VNS)算法都可以用于解决飞行冲突,但PSO算法接近最优解时收敛速度降低,VNS算法的全局搜索能力较差。为融合PSO算法全局搜索的快速收敛特性和VNS算法的局部搜索能力,提出了变邻域搜索改进的粒子群优化算法。仿真结果证明该算法能够快速搜索到全局最优解,继承了二者的优势,同时提高了最终解脱航迹的适应值,并减少了收敛时间。Flight conflict resolution is the key to the safe operation of aircrafts. Both Particle Swarm Optimization （PSO） and Variable Neighborhood Search （VNS） can be used to resolve flight conflict problems. But the rate of PSO will be lower when the solution is close to the optimal one, while VNS is weak at global search. In order to combine fast convergence for global search of PSO and local search of VNS, we proposed enhanced PSO with VNS. The simula- tion results prove that the improved algorithm can search the global optimal solution effectively, meanwhile inherit ad- vantages of both two algorithms and improve the quality of final solved flight tracks.

关 键 词：粒子群优化算法 变邻域搜索算法 飞行冲突解脱 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] V355[自动化与计算机技术—计算机科学与技术]