Evolving hierarchical and recursive teleo-reactive programs through genetic programming

Abstract

Teleo-reactive programs and the triple tower architecture have been proposed as a framework for linking perception and action in agents. The triple tower architecture continually updates the agent's knowledge of the world and evokes actions according to teleo-reactive control structures. This paper uses block stacking problems to demonstrate how genetic programming may be used to evolve hierarchical and recursive teleo-reactive programs. © Springer-Verlag Berlin Heidelberg 2003.