Abstract

We have applied an information theory approach in order to study the problem of convergence to point-like or extended attractors in reaction-diffusion systems. A distance between two states based on the Kullback-Leibler relative information was defined. Different forms of the probability distribution, same of them based on the knowledge of the nonequilibrium potential when accesible, give the possibility to look for a faster and/or more accurate convergence. This approach offers the chance to estimate the am-action basins of the different attractors as well as detecting limit circles, together with an easy evaluation of their periods. (C) 1999 Elsevier Science B.V. All rights reserved.