Regensburg 2010 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 30: Posters II

DY 30.4: Poster

Donnerstag, 25. März 2010, 16:00–18:00, Poster C

Using an evolutionary algorithm to obtain Boolean networks with reliable and robust trajectories — •Christoph Schmal^{1, 2}, Tiago Peixoto¹, and Barbara Drossel¹ — ¹Institut für Festkörperphysik, TU-Darmstadt — ²Fakultät für Physik, Universität Bielefeld

We investigate Boolean networks that follow a given reliable trajectory in state space, which is insensitive with respect to the updating schedules, as recently introduced in [1]. These trajectories have the property that two successive states differ only by the value of one node. We evaluate the robustness of these networks under small perturbations of the dynamics. Here, robustness is defined as the probability that the dynamics return to the reliable trajectory after a perturbation of the state of a single node. In order to achieve higher robustness, we explore the space of possible update functions by using an evolutionary algorithm. With this procedure, we obtain networks that are robust against noise in the update schedule as well as against perturbations of the Boolean values. We compare properties such as the probability distribution of the different types of update functions, the size of the basin of attraction of the reliable trajectory, the transient times and the attractor lengths, for networks before and after running the evolutionary optimization process.

References:

[1] Tiago P. Peixoto and Barbara Drossel, Phys. Rev. E 80, 056102 (2009)