We present a multi-robot model suitable for study of interactions and emergence of rational behavior. We focus on a grammatical approach, and to demonstrate its advantages, we design a model of an adaptive multi-robot community in terms of eco-grammar systems. We show that this grammatical model, based on the blackbord architecture, can naturally involve reinforcement collective learning. We test two learning algorithms in a common environment with almost reactive co-operating robots. Experimental results show that using the grammatical model, the robot community can be successfully trained to find a close-to-optimal solution to a given NP-complete task of a truss construction.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
Two special cases of eco-grammar systems are studied, namely the systems with identical agents, called monocultures and the systems with homogeneous environment characterized by a unary alphabet. The generative power of these eco-grammar systems, given by the languages of the environments, is discussed. Introduced special cases restrict the generative power of eco-grammar systems. Hierarchy or incomparability of language classes with respect to the degree of eco-grammar systems, given by the number of the agents, is shown for various cases of eco-grammar systems.
3
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
Eco-grammar (EG) systems are proposed as a suitable formal framework for the study of some of the computationally relevant properties of the behavior of collections of embodied agents sharing a common environment and acting in it in simple ways. It is illustrated that the computational power of such systems goes - in certain situations - beyond the traditional limits of the Turing-computability.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.