摘要
针对遗传算法应用于移动机器人路径规划时存在早熟和收敛慢的问题,提出了一种改进的遗传算法。通过改良初始种群的生成方法,将八叉树与引导函数结合,提高了初始种群的质量;选择操作中采用轮盘赌和精英保留策略相结合的方式,避免了精英个体的丢失而使算法陷入局部最优;在遗传操作中增加修正算子,在不影响进化的同时保证进化后个体的有效性。仿真实验表明,改进遗传算法相比于基本遗传算法,收敛快,寻优效率高,且不易陷入局部最优,整体性能优于基本遗传算法。
Aiming at the exsiting problems of precocity and slow convergence when the Genetic Algorithm is applied to mobile robot path planning,an improved Genetic Algorithm is proposed.By improving the generation method of the initial population,the octree is combined with the bootstrap function to improve the quality of the initial population.The combination method of roulette and elite retention strategy is used in the selection operation to avoid the algorithm falling into the local optimum due to the loss of the excellent individuals,and the correction operator is added in the genetic operation to ensure the the individual validity after evolution while not affecting the evolution.Simulation experiments show that the improved Genetic Algorithm,compared with the basic Genetic Algorithm,converges quickly,has high efficiency infinding the optimum,and is not easy to fall into the local optimum.It is better than the basic Genetic Algorithm in the overall performance.
作者
李忠林
贾玉婷
蒋晓丽
LI Zhonglin;JIA Yuting;JIANG Xiaoli(Software Engineering Institute of Guangzhou,Guangzhou 510990,China)
出处
《现代信息科技》
2024年第23期180-183,188,共5页
Modern Information Technology
基金
广州软件学院2022年度校级科研项目(ky202203,ky202208)。
关键词
移动机器人
路径规划
遗传算法
改进遗传算法
mobile robot
path planning
Genetic Algorithm
improved Genetic Algorithm
