一种改进的双种群遗传算法在服务机器人仿真中的应用  

Application of an Improved Double Population Genetic Algorithm in Service Robot Simulation

作　　者：刘璇 陈万米[1] 钟灿灿 

机构地区：[1]上海大学机电工程与自动化学院,上海200072

出　　处：《工业控制计算机》2021年第6期62-63,共2页Industrial Control Computer

摘　　要：服务机器人仿真主要研究基于真实场景中服务机器人在完成多个任务时的最优行动路径和行动序列规划。采用一种新的双种群遗传算法,两个子种群的交叉、变异以及复制操作互不相同,同时进化两个群体,其中之一注重于局部搜索,另一个注重于全局搜索,在两个种群之间进行移民来均衡算法的局部和全局能力。实验结果表明:改进后的双种群遗传算法(DPGA)优于传统的服务机器人规划算法。The simulation of service robots mainly studies the optimal action path and action sequence planning when the service robot completes multiple tasks in the real scene.In this paper,a new two population genetic algorithm is proposed.The crossover,mutation and replication operations of two subpopulations are different from each other.Two populations are evolved at the same time,one of which focuses on local search and the other on global exploration.Migration between the two populations is carried out to balance the local and global capabilities of the algorithm.The experimental results show that the improved DPGA is superior to the traditional service robot planning algorithm.

关 键 词：多种群遗传算法 服务机器人仿真 动作序列规划 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP242[自动化与计算机技术—控制科学与工程]