摘要
针对遗传类算法收敛精度差和收敛速度慢等问题,本文将育种系统的管理运作思想引入遗传算法,构造了一种新的全局优化算法—育种算法。通过对搜索和进化操作过程进行分析,指出了算法收敛到全局最优的途径和方法,提出了利用简单的随机采样实现全局搜索和采用基因置换技术实现交叉进化的思想策略,建立了算法模型并确定了相应的控制参数和终止准则。实验表明,该算法能够实现精确搜索并实现计算精度和成本之间的平衡,可以避免遗传算法的早熟收敛问题和大量的冗余运算,提高了优化计算的速度和可靠性。
Introducing the management idea of breeding systems into the genetic algorithm, a novel optimization algorithm identified breeding algorithm (BA) is originated to overcome the problems existing in genetic algorithms such as inferior precision and low speed of convergence. Through analysis to the process of search and evolution, the access to the global optimization is illuminated. The strategy is proposed, by which the exploration is reached through random sampling, and the exploitation is achieved by the technology of gene replacement. The computational modal is constructed, and the related control parameters and termination criterion are defined. The testing results show that BA can undertake the task of precise search and achieve the balance between computational precision and cost. It can successfully solve the problems of premature convergence and large number of redundant computations existing in conventional genetic algorithms, and hence the reliability and speed of optimization computation can be greatly promoted.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2007年第1期82-86,共5页
Systems Engineering and Electronics
关键词
遗传算法
交叉育种
基因置换
收敛性
数值实验
genetic algorithm
cross breeding
gene replacement
convergence performance
computation precision