基于约束支配多目标进化算法的航路改航方法  

Rerouting method based on constrained-dominate multi-objective evolutionary algorithm

作　　者：毛芸苹[1] 杨红雨[1] 

机构地区：[1]四川大学计算机学院国家空管自动化系统技术重点实验室

出　　处：《计算机工程与设计》2013年第3期1008-1012,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(60736046)

摘　　要：针对具有多目标多约束特征的航路改航问题,提出了基于约束支配的非支配排序多目标进化算法(NSGA-II)的路径规划方法。采用了约束支配规则处理空中交通管制规则中关于改航的多个约束条件,建立个体解之间的约束支配关系。在产生初始群体时进行分区产生随机的浮点坐标,并提前筛选出符合航向角改变量约束的可接受解,增加删除算子用以处理不可接受的不可行解,进化算法迭代完成后,建立优化算子用以优化迭代产生的路径。在块状和离散状飞行限制区条件下分别进行仿真实验,实验结果表明,该方法能产生符合目标函数优化且路径复杂度较小的最优解。The path planning method which using constraineddonminate multiobjective evolutionary algorithm is applicated for the multiobjective and multiconstrained rerouting. The constraints of the rerouting which obey the air traffic rules are dealt with the constraineddonminate principle, the constraineddonminate relations between the individuals is built at the same time. First the random float coordinates of the initial generation are generated in the separated areas, and the acceptable solutions are screened in advance which meet the constraint on the change of the angular size, a deleted operator is added to deal with unac ceptable and infeasible solutions. After the iterations completed, a optimal operator is added to optimize the path. Finally, simu lations are done under the massive and discrete flight restricted conditions, results show that a set of solutions with smaller order of complrxity are generated and they meet the objective functions＇ optimization.

关 键 词：约束支配 多目标 进化算法 改航 空中交通管制 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]