Local Search Methods for Tree Chromosome Structure in a GA to Identify Functions
Mitsukuni Matayoshi, Morikazu Nakamura, Hayao Miyagi · IEEJ Transactions on Electronics Information and Systems · 2006
In this paper, Local search methods for “Tree Chromosome Structure in a Genetic Algorithm to Identify Functions" which succeeds in function identifications are proposed. The proposed method aims at the identification success rate improvement and shortening identification time. The target functions of identification are composed of algebraic functions, primary transcendental functions, time series functions include a chaos function, and user-defined one-variable funcions. In testing, Kepler's the third law is added to Matayoshi's test functions(7)-(9). When some functions are identified, the improvement of identification rate and shortening time are indicated. However, we also report some ineffectual results, and give considerations.