Abstract
We present symmetric and asymmetric similarity measures for labeled directed rooted graphs that are inspired by the simulation and bisimulation relations on labeled transition systems. Computation of the similarity measures has close connections to discounted Markov decision processes in the asymmetric case and to perfect-information stochastic games in the symmetric case. For the symmetric case, we also give a polynomial-time algorithm that approximates the similarity to any desired precision. © Springer-Verlag Berlin Heidelberg 2006.
Cite
CITATION STYLE
Sokolsky, O., Kannan, S., & Lee, I. (2006). Simulation-based graph similarity. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3920 LNCS, pp. 426–440). https://doi.org/10.1007/11691372_28
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
