摘要
针对传统遗传算法进行路径规划时搜索空间大、出现过多搜索冗余和收敛效率低等问题,提出在基于网格的遗传算法上加入弹性网格概念。在低密度的网格地图下求解当前最优路径,针对转向点局部增加网格密度,进一步路径寻优,如此重复,以减小算法搜索空间,提高路径规划效率;同时,给出自适应变异概率,使其根据各代路径离散程度自适应调整大小,以提高各代路径多样化,并进行仿真分析和试验。仿真结果表明:平均迭代次数明显少于传统遗传算法,收敛速度得到改善,最终寻优路径达到与障碍物无干涉,总长度明显较短的基本预期效果。
The variable mesh method is introduced into the genetic algorithm to improve the efficiency of optimal path searching of unmanned surface vehicles.The concept is to search a generally optimal path in low mesh density and then refine it by further searching the optimal tuning points in higher mesh density.The process is repeated until the results converge.The searching space is reduced and the efficiency is improved.The adaptive probability of mutation that changes according to the dispersion of search results is used for diversity.Simulations are performed,which shows that the algorithm is good in avoiding obstacles and gives much shorter paths with noticeably less iterations and better convergence efficiency than the basic genetic algorithm.
作者
余文曌
佘航宇
欧阳子路
YU Wenzhao;SHE Hangyu;OUYANG Zilu(School of Transportation, Wuhan University of Technology,Wuhan 430063, China;School of Logistics Engineering, Wuhan University of Technology,Wuhan 430063, China;State Key Laboratory of Ocean Engineering, Ministry of Education,Shanghai 200240, China;Key Laboratory of Marine Intelligent Equipment and System, Ministry of Education, Shanghai 200240, China)
出处
《中国航海》
CSCD
北大核心
2018年第4期101-105,共5页
Navigation of China
基金
中央高校基本科研业务费专项资金(20171049702005)
武汉理工大学国家级大学生创新创业训练计划
关键词
路径规划
遗传算法
弹性网格
自适应变异函数
path planning
genetic algorithm
flexible mesh
adaptive mutate function
