机器人足球比赛系统中的机器人任务分配方法  

Robot Task Allocation Method in Robot Soccer Match System

作　　者：王兴[1] 杨宜民[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《计算机与现代化》2012年第2期14-18,共5页Computer and Modernization

摘　　要：针对基于遗传算法的多机器人任务分配方法中,由于初始种群是随机产生不能很好地表征整个解空间,而容易陷入局部最优解的问题,提出一种新的多机器人任务分配方法。该方法根据机器人效用函数值来确定个体基因,从而产生初始种群,并引进分层遗传算法来实现具有不同最优基因的群体分开演化。仿真实验表明,该方法比传统的遗传算法有更高的寻优效率和更快的收敛速度。According to the problem in multi-robot task allocation method based on genetic algorithm,that the initial population generated randomly is not well characterized the whole solution space,and it＇s easy to fall into local optimal solution,this paper proposes a new multi-robot task allocation method.This method is based on the utility value of the robot to determine the individual＇s genes,resulting in the initial population,and by hierarchical genetic algorithm different groups with different optimal genes are evolution separately.Simulation results show that this method has higher efficiency and faster convergence.

关 键 词：任务分配 分层遗传算法 初始种群 效用函数值 最优基因 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]