In this paper we present our preliminary results of the application of modified Cellular Automata (CA) used as a phenomenological models of some processes that evolve in time and in space. They will be called Cellular-Like Automata (CLA). The rule f that defines the state of a cell at the following steps is defined by means of an Artificial Neural Network (ANN) with hidden layers. The CA model is obtained by a usual training of the ANN. The training data for the network are made of the observations of the process we want to approximate. We check the feasibility of the proposed approach by modelling the well known Convey’s game of life (both 1D and 2D patterns) and a problem of temperature propagation.
Cellular Automata with learning that identify a behaviour governed by an unknown rule
BOSO, DANIELA;
2013
Abstract
In this paper we present our preliminary results of the application of modified Cellular Automata (CA) used as a phenomenological models of some processes that evolve in time and in space. They will be called Cellular-Like Automata (CLA). The rule f that defines the state of a cell at the following steps is defined by means of an Artificial Neural Network (ANN) with hidden layers. The CA model is obtained by a usual training of the ANN. The training data for the network are made of the observations of the process we want to approximate. We check the feasibility of the proposed approach by modelling the well known Convey’s game of life (both 1D and 2D patterns) and a problem of temperature propagation.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.