This paper is a collection of previous studies for function identification by simple genetic algorithm (GA) [1] with tree chromosome structure which has been proposed in [2]-[7], and gives the details more than survey...
详细信息
ISBN:
(纸本)9781643681351;9781643681344
This paper is a collection of previous studies for function identification by simple genetic algorithm (GA) [1] with tree chromosome structure which has been proposed in [2]-[7], and gives the details more than survey paper. This paper also aims to introduce the studies which were written in Japanese. In this paper, there are five main points. First, a tree chromosome structure, which is the core idea of the studies, is introduced. The tree chromosome structure makes GA succeed in function identification called symbolic regression. Second, the proposed GA with tree chromosome structure succeeded in identifying the target functions from the observed data are shown indeed. The target functions are algebraic functions, primary transcendental functions, time series functions including chaos function, and user-defined one-variable functions. Third, to find function represented with some parentheses, a hierarchical tree chromosome structure is introduced. Forth, some local search methods to aim at the improvement for identification success rate and shortening identification time are introduced. In the end of this paper, the proposed tree and hierarchical tree chromosome structure can be adapted for identifying Boolean functions are laid out.
暂无评论