In this paper we study the properties of undirected vertex-labeled graphs and the limitations on the languages that they represent. As a main result of this paper we define the necessary and sufficient conditions for the languages to be representable by a class of undirected vertex-labeled graphs and its subclasses. We assume that all obtained results and techniques are transferable to the case of undirected edge-labeled graphs and might give us similar results. The simplicity of necessary conditions emphasizes the naturalness of the result. The proof of their sufficiency is quite non-trivial and it is based on a new notion of quasi-equivalence, that is significantly different from Myhill-Nerode equivalence and might not be reduced to it.
[1]
Tod S. Levitt,et al.
Qualitative Navigation for Mobile Robots
,
1990,
Artif. Intell..
[2]
Samuel Eilenberg,et al.
Automata, languages, and machines. A
,
1974,
Pure and applied mathematics.
[3]
Michael R. M. Jenkin,et al.
Map validation and robot self-location in a graph-like world
,
1997,
Robotics Auton. Syst..
[4]
G. Kilibarda,et al.
On behaviour of automata in labyrinths
,
1993
.
[5]
Matthias Jantzen,et al.
Refining the hierarchy of blind multicounter languages and twist-closed trios
,
2003,
Inf. Comput..