基于线性变换的变异算子及在机器人进化计算中的应用  

Adaptive Mutation Operator Based on Linear Transformation and Its Application to Evolutionary Computation of Robot

作　　者：彭伟民[1] 

机构地区：[1]广东海洋大学信息学院计算机系,湛江524088

出　　处：《现代计算机》2007年第4期97-98,共2页Modern Computer

摘　　要：进化算子对进化计算行为和性能有直接影响。根据进化阶段的不同和个体适应度的大小,采用线性变换法建立能保持群体多样性和收敛性的自适应的变异算子的表达式,并在Evorobot系统中仿真实现。仿真结果表明,基于线性变换的自适应的变异算子改进了进化计算性能,体现了优越性。Evolutionary operators have a direct impact on behavior and performance of evolutionary computation. Based on different stages of evolution and the size of individual fitness, this paper establishes a type of self-adaptive mutation operator with linear transformation which can maintain diversity and convergence of groups. Then the related simulation experiments are implemented on platform Evorobot. Simulation results show that the self-adaptive mutation operator based on linear transformation improves the performance of evolutionary computation and reflects the superiority.

关 键 词：进化计算 线性变换 变异算子 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TP18[自动化与计算机技术—控制科学与工程]