This paper presents Genetic Algorithm based Cellular Automata for predicting chaotic time series. The simple and interesting Tent map,x n+1=a(1-2|1/2-x n), is used here. The values are first coded into binary form and the first N of them are given as initial condition to Cellular Automata (CA). Genetic Algorithm evolves CA rules, and the second N values of the map are the CA output. One of the important advantages of this work in compare with others is its ability for parallel computing.
Foroogh Motallebzadeh , Caro Lucas
Electrical and Computing Engineering Department, Tehran University, Tehran, Iran
School of Cognitive Sciences, Institute for studies in theoretical Physics and Mathematics (IPM), Tehran