摘要
研究了遗传算法中的位重要性和位收敛顺序性,给出了重要位、模式、参数区间和目标函数四者之间的关系,提出了一种新的进化算法———位重要性进化算法(Bit Importance Evolutionary Algorithm,BIEA).BIEA通过检测组成个体各位的重要性,对于重要位,加快其收敛;对于非重要位,保持其多样性.数据实验表明:BIEA在收敛速度上要优于遗传算法,同时BIEA也可以有效地解决一类遗传算法很难解决的强欺骗性问题.
The bit importance and convergent order of bit in genetic algorithms (GA) are studied. Relationships among important bits, schema, parameter range and target function are presented,and then a novel evolutionary algorithm, called Bit Importance Evolutionary Algorithm (BIEA) is proposed. BIEA detects the important bits in a chromosome at first, and then speeds up the convergence of important bits, while maintaining the diversity of unimportant bits. Numerical experiments show that compared with GA, BIEA has a better convergent velocity and can solve some hard deceptive problems which can't be solved effectively by GA.
出处
《计算机学报》
EI
CSCD
北大核心
2006年第6期992-997,共6页
Chinese Journal of Computers
基金
国家自然科学基金(60271033)资助
关键词
遗传算法
概率分布估计算法
位重要性
模式
位重要性进化算法
genetic algorithms
estimation of distribution algorithms
bit importance
schema
bit importance evolutionary algorithm