Optimizing Cellular Automata through a Meta-model Assisted Memetic Algorithm

Donato D’Ambrosio¹, Rocco Rongo², William Spataro¹, and Giuseppe A. Trunfio³

¹Department of Mathematics, University of Calabria, 87036, Rende, CS, Italy

²Department of Earth Sciences, University of Calabria, 87036, Rende, CS, Italy

³DADU, University of Sassari, 07041, Alghero, SS, Italy

Abstract. This paper investigates the advantages provided by a Meta-model Assisted Memetic Algorithm (MAMA) for the calibration of a Cellular Automata (CA) model. The proposed approach is based on the synergy between a global meta-model, based on a radial basis function network, and a local quadratic approximation of the fitness landscape. The calibration exercise presented here refers to SCIARA, a well-established CA for the simulation of lava flows. Compared with a standard Genetic Algorithm, the adopted MAMA provided much better results within the assigned computational budget.

Keywords: Cellular Automata, Model Calibration, Meta-modelling, Memetic Algorithms

LNCS 7492, p. 317 ff.

Full article in PDF | BibTeX

lncs@springer.com
© Springer-Verlag Berlin Heidelberg 2012