On a King Crossover Based Elitist Genetic Algorithm
For more than twenty years, in classical elitist Genetic Algorithm (EGA), the only function of the reserved elitist is to reserve the best solution that was found by the algorithm so far. Inspired by the fact that the king of some animal kingdoms has privilege in all aspects of animal life including mating and propagation, we proposed a king crossover strategy. And combining EGA with king crossover strategy, we developed a kind of new king-crossover based elitist Genetic Algorithm (KEGA). This paper studies KEGA in more detail. The computational overhead of KEGA, the effect of KEGA on diversity of population and the effect of king crossover probability on convergence performance were studied with a modified F6 function, and gives out some new concludes. The empirical results show that KEGA can improve performance significantly and the computational overhead is negligible.
Rirong Zheng Ronghua Xu Ya-Nan Song
College of Automation Guangdong University of Technology Guangzhou, 510090, China
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）