Fuzzy Behavior-Based Control. 2nd Report, Learning with a Virus-Evolutionary Genetic Algorithm with Species.
Keigo Watanabe, Kiyotaka Izumi, Takehide MIYAZAKI · TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C · 2000
In this paper, we describe a learning of fuzzy behavior-based control system by applying a virusevolutionary genetic algorithm with species. We first state the development of a virus-evolutionary genetic algorithm with species(VEGAS), in which one population is divided into some subpopulations(i.e., species)so that a virus infection between species can be introduced as well as a virus infection within a species. We then apply the VEGAS to the learning of a fuzzy behavior-based control system for a mobile robot with two independent driving wheels, where a control problem of acquiring multi-objective points is considered. The effectiveness of the present learning method is demonstrated by a set of computer simulations, compared to the conventional method with a simple GA.