この論文を読む／探す
抄録
A new evolutionary model named "Genetic Network Programming, GNP" has been proposed. GNP represents its solutions as network structures, which realizes better expression ability than GA and GP which use string and tree structures, respectively. GA, GP and the conventional GNP are based on offline learning, so it is difficult to adapt to the dynamical environments because offline learning methods cannot change their solutions until one generation ends. In order to adapt to dynamical environments quickly, the basic studies of online learning of GNP have been done. In this paper, GNP Learning and Evolution is proposed in order to acquire to explore the wide space of solutions using offline learning effectively avoiding ineffective trials.
