带动态障碍区的自由区域路径实时优化问题的混合算法  

A Hybrid Algorithm for Real-time Optimization of Path in Free Area with Dynamic Obstacle Area

作　　者：徐里 丁炜[2] 施进[2] 张海林[3] 胡小兵[4] 

机构地区：[1]北京师范大学地表过程与资源生态国家重点实验室,北京100875 [2]浙江省标准化研究院,浙江杭州310006 [3]交通运输部公路科学研究院,北京100088 [4]北京师范大学减灾与应急管理研究院,北京100875

出　　处：《公路交通科技》2017年第11期116-125,共10页Journal of Highway and Transportation Research and Development

摘　　要：自由区域路径优化问题可以在除障碍区之外的整个区域内自由规划路径,为了解决带动态障碍区的自由区域路径实时优化问题,提出了一种遗传算法(GA)加滑动地平线控制(RHC)的混合算法。首先,建立和讨论了带动态障碍区的自由区域路径实时优化问题的数学模型。然后,详细描述了提出的遗传算法加滑动地平线策略的混合算法,阐述了混合算法中滑动地平线控制与遗传算法结合的关键步骤之一:可变长度染色体的设计。全面探讨了混合算法中滑动地平线长度的选择对于混合算法的影响,说明了滑动地平线控制策略中不同终端加权设计的路径优化效果,并通过终端加权的设计,以保证路径规划的可行性和优化性能。仿真结果表明,遗传算法(GA)加滑动地平线控制(RHC)的混合算法非常有效,在确定性的环境条件下,获得与现有GA算法几乎相同的求解性能,而在动态和不确定的环境下,新算法则取得了更佳的求解效果。在这两种情况下,带RHC的混合算法的在线计算时间是单纯GA算法的一小部分。The issue of path optimization of free area can be made freely in the entire area except the obstacle area. In order to solve the issue of real-time path optimization in free area with dynamic obstacles, a hybrid algorithm integrating Genetic Algorithm （GA） with Receding Horizon Control （RHC） is proposed. To begin with, a mathematical model for the issue of real-time path optimization in free area with dynamic obstacles is established and discussed. Furthermore, the details of the proposed hybrid algorithm based on GA with RHC is described. One of the key steps in the combination between RHC and GA in the hybrid algorithm is described： the design of variable length chromosomes. The influence of the selection of the receding horizon length on the hybrid algorithm is discussed comprehensively. Then, the path optimization effect of different terminal weighting designs in RHC strategy is illustrated, and the design of the terminal weighting is performed to ensure that the route planning is feasible and optimal. The simulation result illustrates that （1）the hybrid algorithm that combining GA and RHC is very efficient, it works better in a dynamic and uncertain environment than in the certain environment when achieving almost the same optimal solution as an existing GA algorithm. In both cases ; （2） the online computing time of the hybrid algorithm with RHC is a small part of the pure GA algorithm.

关 键 词：交通工程 滑动地平线控制 混合算法 路径实时优化 自由区域 动态障碍 

分 类 号：U491[交通运输工程—交通运输规划与管理]