摘要： 针对改进族群进化算法(EGEA)中对族群模式的有效采样问题，提出以竞争指数为选择指标的常规选择、以族群为选择单位的常规选择以及基于竞争指数的模拟退火排序选择3种选择机制，并通过多维函数优化问题进行实验分析，结果表明3种选择机制都能够在一定程度上提高EGEA的搜索效率，且基于竞争指数的模拟退火排序选择算子的实现更简便，对EGEA综合性能的提高也更明显。
关键词:
族群进化算法,
选择算子,
模拟退火排序选择
Abstract: In order to improve the sampling efficiency of ethnic group schemas in Ethnic Group Evolution Algorithm(EGEA), three kinds of selection mechanism, including race exponent based conventional selection, race exponent based ethnic group selection and race exponent based annealing rank selection, are proposed. Simulations of numerical optimization show race exponent based annealing rank selection mode is easy to implement and reduce the computational effort greatly, which is more suitable for EGEA.
Key words:
Ethnic Group Evolution Algorithm(EGEA),
selection operator,
simulated annealing rank selection
中图分类号:
陈 皓;崔杜武. 族群进化算法的选择机制[J]. 计算机工程, 2010, 36(3): 200-202.
CHEN Hao; CUI Du-wu. Selection Mechanism of Ethnic Group Evolution Algorithm[J]. Computer Engineering, 2010, 36(3): 200-202.