摘要
针对传统遗传算法在解决采样点模型下的路径规划问题中存在种群进化速度慢、最终得到的最优路径偏长的问题,提出一种改进遗传算法.为提高初始种群的质量,提出自适应调整步长来限制子节点的选择范围,在这个范围内随机选取两条父代路径,在对应两节点之间形成一个矩形的子代节点搜索范围,每个范围中选取一点并依次连接得到交叉后的子代个体,避免了采样点模型下交叉点不足导致交叉无效的问题;为解决变异效果不可控的问题,以变异父节点的前后两点连线为引导,越靠近该连线的节点被选为变异子节点的概率越大,使得在变异点的选择更具有方向性.对比实验结果表明:所提出的改进遗传算法当处理基于采样点路径规划问题时可以有效提高寻路效率,最优路径收敛速度比传统遗传算法提高约60%,最优路径长度最多减少了2.42 m,比其他文献算法的最优路径收敛速度最多提高56%.
A modified genetic algorithm was proposed to address the issues of slow population evolution,and longer optimal path obtained by traditional genetic algorithms when solving path planning problems under the sampling point model.To enhance the quality of the initial population,adaptive adjustment of step size was suggested to limit the selection range of offspring nodes,within which random selection was conducted.Two parent paths were randomly selected,forming a rectangular search area for offspring nodes between corresponding pairs of nodes.A point was chosen from each area,and connections were made sequentially to obtain offspring individuals after crossover,thereby avoiding ineffective crossovers due to insufficient intersection points in the sampling point model.To address the issue of unpredictable mutation effects,the line connecting the preceding and succeeding points of the selected mutation parent node served as a guide.The closer a node was to this line,the higher the probability of it being selected as the mutation offspring node,rendering the mutation point selection more directional.Comparative experiment results show that the proposed modified genetic algorithm effectively improves pathfinding efficiency when dealing with path planning problems based on sampling points.The convergence speed of the optimal path using the modified algorithm is approximately 60%higher than that of traditional genetic algorithms,and the length of the optimal path is reduced by up to 2.42 m,with the highest improvement in the convergence speed of the optimal path reaching 56%compared to other literature algorithms.
作者
王雷
王艺璇
李东东
王天成
WANG Lei;WANG Yixuan;LI Dongdong;WANG Tiancheng(School of Mechanical and Automotive Engineering,Anhui Polytechnic University,Wuhu 241000,Anhui China)
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2024年第5期158-164,共7页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
安徽省高校优秀拔尖人才培育项目(gxbjZD2022023)
安徽省高校自然科学研究重点资助项目(2022AH050978,2023AH050935,2023AH052915)
芜湖市科技计划资助项目(2022jc26)
安徽工程大学检测技术与节能装置安徽省重点实验室开放研究基金资助项目(JCKJ2021A06,JCKJ2022B01)
安徽工程大学-鸠江区产业协同创新专项基金资助项目(2022cyxtb6)
安徽工程大学科研启动基金资助项目(2022YQQ002).
