This paper presents an extended genetic algorithm based on the neutral theory of molecular evolution. The motivation for the extension originates from the consideration that premature convergence cannot be avoided as ong as a GA employs neo-Darwinism as its framework. The proposed GA, which is called operon-GA, has new genetic mechanisms inspired by the operon structure found in DNA sequences of microbial organisms, so that individuals can develop simple but effective genetic regualtory systems for adaptive genetic search and autonomous diversity control. Several computer simulations were conducted to discuss the evolving behaviors and optimization performance using problems in-cluding different types of GA-difficulties.
展开▼