摘要
在机器人进化计算中,建立一个既能保持群体多样性与收敛性,又能提高进化速度与性能的自适应性适应度函数表达式,对于提高机器人的进化水平尤为重要。采用指数变换法构造适应度函数模型,并在Evorobot系统中仿真实现。仿真结果表明,基于指数变换的适应度函数更具优越性。
In the evolutionary computation of robot, it is especially important to improve the evolution level to establish a self-adaptive fitness function expression, which can keep the groups’ diversity and convergence and enhance the evolution speed and performance. This paper proposes a mathematical model for fitness function with index transformation. Then the related simulation experiments are implemented on platform Evorobot. The simulation results indicate that the fitness function based on index transformation is more advantageous.
出处
《计算机与数字工程》
2007年第7期73-74,共2页
Computer & Digital Engineering
关键词
进化计算
指数变换
适应度函数
evolutionary computation, index transformation, fitness function
